NIST NCSTAR 1-7
Federal Building and Fire Safety Investigation of the
World Trade Center Disaster
Occupant Behavior, Egress, and
Emergency Communications
Jason D. Averill
Dennis S. Mileti
Richard D. Peacock
Erica D. Kuligowski
Norman Groner
Guylene Proulx
Paul A. Reneke
Harold E. Nelson
nist
National Institute of Standards and Technology • Technology Administration • U.S. Deportment of Commerce
NIST NCSTAR 1-7
Federal Building and Fire Safety Investigation of the
World Trade Center Disaster
Occupant Behavior, Egress, and
Emergency Communications
Jason D. Averill Norman Groner
Building and Fire Research Laboratory John Jay College
National Institute of Standards and Technology
Guylene Proulx
Dennis S. Mileti National Research Council Canada
University of Colorado - Boulder
Paul A. Reneke
Richard D. Peacock Building and Fire Research Laboratory
Erica D. Kuligowski National Institute of Standards and Technology
Building and Fire Research Laboratory
National Institute of Standards and Technology Harold E. Nelson
Consultant
September 2005
^ ,
**TES O* **
U.S. Department of Commerce
Carlos M. Gutierrez, Secretary
Technology Administration
Michelle O'Neill, Acting Under Secretary for Technology
National Institute of Standards and Technology
William Jeffrey, Director
Disclaimer No. 1
Certain commercial entities, equipment, products, or materials are identified in this document in order to describe a
procedure or concept adequately or to trace the history of the procedures and practices used. Such identification is
not intended to imply recommendation, endorsement, or implication that the entities, products, materials, or
equipment are necessarily the best available for the purpose. Nor does such identification imply a finding of fault or
negligence by the National Institute of Standards and Technology.
Disclaimer No. 2
The policy of NIST is to use the International System of Units (metric units) in all publications. In this document,
however, units are presented in metric units or the inch-pound system, whichever is prevalent in the discipline.
Disclaimer No. 3
Pursuant to section 7 of the National Construction Safety Team Act, the NIST Director has determined that certain
evidence received by NIST in the course of this Investigation is "voluntarily provided safety-related information" that is
"not directly related to the building failure being investigated" and that "disclosure of that information would inhibit the
voluntary provision of that type of information" (15 USC 7306c).
In addition, a substantial portion of the evidence collected by NIST in the course of the Investigation has been
provided to NIST under nondisclosure agreements.
Disclaimer No. 4
NIST takes no position as to whether the design or construction of a WTC building was compliant with any code
since, due to the destruction of the WTC buildings, NIST could not verify the actual (or as-built) construction, the
properties and condition of the materials used, or changes to the original construction made over the life of the
buildings. In addition, NIST could not verify the interpretations of codes used by applicable authorities in determining
compliance when implementing building codes. Where an Investigation report states whether a system was
designed or installed as required by a code provision, NIST has documentary or anecdotal evidence indicating
whether the requirement was met, or NIST has independently conducted tests or analyses indicating whether the
requirement was met.
Use in Legal Proceedings
No part of any report resulting from a NIST investigation into a structural failure or from an investigation under the
National Construction Safety Team Act may be used in any suit or action for damages arising out of any matter
mentioned in such report (15 USC 281a; as amended by P.L. 107-231).
National Institute of Standards and Technology National Construction Safety Team Act Report 1-7
Natl. Inst. Stand. Technol. Natl. Constr. Sfty. Tm. Act Rpt. 1-7, 298 pages (September 2005)
CODEN: NSPUE2
U.S. GOVERNMENT PRINTING OFFICE
WASHINGTON: 2005
For sale by the Superintendent of Documents, U.S. Government Printing Office
Internet: bookstore.gpo.gov — Phone: (202) 512-1800 — Fax: (202) 512-2250
Mail: Stop SSOP, Washington, DC 20402-0001
Abstract
This report describes the occupant evacuation of World Trade Center (WTC) 1 and WTC 2 on
September 11, 2001. Multiple sources of information were collected and analyzed: over 1,000 new
interviews with survivors (including 803 telephone interviews, 225 face-to-face interviews, and 5 focus
groups); over 700 published interviews; 9-1-1 emergency calls; transcripts of emergency
communications, historical building design drawings, memoranda, and calculations; formal complaints
filed with the Occupational Safety and Health Administration; and other relevant materials.
The egress system, including stairwells and elevators, was described and compared to requirement of both
contemporary and current code requirements. This report documents the emergency procedures, both as
they were designed to be implemented, as well as how they were actually implemented on
September 11,2001.
The population in WTC 1 and WTC 2 on September 11, 2001, at 8:46:30 a.m. was enumerated and
described, where the characteristics of the population were relevant to the subsequent evacuation,
including training, experience, mobility status, among others. The progress of the evacuation of both
towers was described in a quasi-chronological manner from 8:46:30 a.m. when WTC 1 was attacked, until
10:28:22 a.m., when WTC 1 collapsed.
Causal models were built to explore the sources of evacuation initiation delay (why people did not
immediately start to leave the building) as well as normalized stairwell evacuation time (how long the
average occupant spent in the stairwells per floor). Issues identified as contributing to either speeding or
aiding the evacuation process were explored. Egress simulations provided context for estimating how
long WTC 1 and WTC 2 would have taken to evacuate with different populations, using different models,
and subject to different damage to the building.
Keywords: Building fires, egress, egress modeling, emergency communication, evacuation, human
behavior, interviews, World Trade Center.
NIST NCSTAR 1-7, WTC Investigation iii
Abstract
This page intentionally left blank
iv NIST NCSTAR 1-7, WTC Investigation
Table of Contents
Abstract iii
List of Figures ix
List of Tables xi
List of Acronyms and Abbreviations xiii
Metric Conversion Table xv
Preface xix
Acknowledgments xxix
Executive Summary xxix
Chapter 1
Background and Introduction 1
1.1 Historical Incidents in Which Egress Was Significant 1
1.2 Previous Fires and Evacuation Incidents in the World Trade Center 4
1.3 Scope of Occupant Behavior, Egress, and Emergency Communications Project for the
NIST World Trade Center Investigation 6
1.4 Report Organization 7
Chapter 2
Design of the World Trade Center Egress System 9
2.1 Overall Building Description 9
2.1.1 Description of the Towers 10
2.2 Building Systems 18
2.2.1 Egress Calculations 18
2.2.2 Stairwells 27
2.2.3 Elevators 32
2.2.4 Emergency Communication System 35
2.3 The Human Element 37
2.3.1 Responsibilities of the Fire Safety Director 38
2.3.2 Responsibilities of the Deputy Fire Safety Director 38
2.3.3 Responsibilities of the Operations Control Center Supervisor 39
2.3.4 Responsibilities of the Operations and Maintenance Management 39
2.3.5 Floor Warden System 40
2.3.6 Occupant and Tenant Training 40
NISTNCSTAR 1-7, WTC Investigation
Table of Contents
2.4 Changes to the Egress System after the 1993 Bombing 41
2.5 Building Code Analysis 43
2.5.1 Egress in the Building Codes 43
2.5.2 New York City Building Code (1968) 45
2.5.3 New York City Building Code (October 2003) 46
2.5.4 International Building Code (2000) 46
2.5.5 NFPA 5000 and NFPA 101 - Life Safety Code (2003) 46
2.5.6 Comparison of Current Code Requirements 47
Chapter 3
First-Person Data Collection and Analysis Methods 49
3.1 Purpose and Scope of Collection of First-Person Data 49
3.2 Methods 49
3.2.1 Published First-Person Accounts 49
3.2.2 Telephone Surveys 50
3.2.3 Face-to-Face Interviews 59
3.2.4 Focus Groups 63
3.2.5 Audio, Video, Photographic, and Records Collection 64
3.3 Analysis of First-Person Data 64
Chapter 4
September 11, 2001, Before the Attacks 67
4.1 Building Population on September 11, 2001 67
4.2 Occupant Characteristics 68
4.3 Previous Experience 70
4.4 Occupant Preparedness 70
Chapter 5
September 11, 2001, 8:46:30 a.m. Flight 11 Crashes into WTC 1 73
5.1 Introduction 73
5.2 Initial Observations in WTC 1 75
Chapter 6
September 11, 2001, 8:47 a.m. - 9:02 a.m. Occupants React to the Attack on WTC 1 79
6.1 Conditions Worsen in WTC 1 79
6.1.1 Activities and Information 81
6.1.2 Emergency Response at the Fire Command Station, Lobby, WTC 1 85
NIST NCSTAR 1-7, WTC Investigation
Table of Contents
6.1.3 Survivors Begin Their Evacuation 86
6.2 Initial Observations and Reactions from WTC 2 89
6.2.1 Activities and Information- WTC 2 91
6.2.2 Survivors Begin Their Evacuation 95
6.2.3 Elevator Use in WTC 2 96
6.2.4 Announcements in WTC 2 98
Chapter 7
September 11, 2001, 9:02:59 a.m. Flight 175 Crashes into WTC 2 99
Chapter 8
September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m. 56 Critical Minutes 105
8.1 Survivors from at or above Floor 78 in WTC 2, After Impact 106
8.2 Progress of Evacuation below the Impact Region 107
8.3 Evacuation Route through Concourse 109
8.4 More Announcements in WTC 2 110
8.5 Conditions Continue to Deteriorate in Both Towers Ill
Chapter 9
September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m. Collapse 115
9.1 Evacuating WTC 1 115
9.2 Location of Victims and Survivors within the Building 119
9.2.1 Fatalities in Elevators 122
Chapter 10
Discussion and Analysis 125
10.1 Overview of Total Evacuation Time 125
10.1.1 Evacuation Initiation Delay Time in WTC 1 and WTC 2 127
10.1.2 Time and Average Travel Speed in the Stairwells, WTC 1 129
10.2 Factors Influencing Total Evacuation Time 130
10.2.1 Predicting Evacuation Delay in World Trade Center 131
10.2.2 Predicting Normalized Stairwell Evacuation Time in WTC 1 on
September 11, 2001 135
10.3 Issues That Impacted Occupant Evacuation 138
10.3.1 Environmental Cues and Information 138
10.3.2 Building Alarm Systems 141
10.3.3 Public Address Announcements 142
10.3.4 9-1-1 System 145
NIST NCSTAR 1-7, WTC Investigation vii
Table of Contents
10.3.5 Previous Evacuation Experience 146
10.3.6 Occupant Activities 147
10.3.7 Aids and Constraints to Evacuation 151
10.3.8 Emergency Responders and Building Authorities 153
10.3.9 Occupant Experience 158
10.3.10 Mobility-Impaired Occupants 158
10.4 Evacuation Simulations 160
10.4.1 Egress Simulation Results 160
10.4.2 Egress Modeling Conclusions 164
Chapter!!
Findings 165
11.1 Building Population and Demographics 165
11.2 Evacuation 165
11.3 Delays in Evacuation 167
11.4 Emergency Management and Planning 168
11.5 Occupants with Mobility Impairments 168
Chapter 12
References 171
Appendix A
Telephone Interview Script 175
Appendix B
Qualitative Analysis Coding 211
Appendix C
Causal Modeling 217
Appendix D
Egress Modeling 237
viii NIST NCSTAR 1-7, WTC Investigation
List of Figures
Figure P-l. The eight projects in the federal building and fire safety investigation of the WTC
disaster xxi
Figure 2-1. WTC site plan 10
Figure 2-2. Typical WTC tower architectural floor plan 11
Figure 2-3. Shopping mall layout underneath WTC plaza 13
Figure 2-4. WTC 1 lobby (concourse) level 14
Figure 2-5. WTC 2 lobby (concourse) level 15
Figure 2-6. Typical WTC tenant spaces 17
Figure 2-7. A WTC 4 trading floor 17
Figure 2-8. 1993 Windows on the World egress calculation, floor 106 25
Figure 2-9. 1993 Windows on the World egress calculation, floor 107 26
Figure 2-10. 44 in. stairwell in WTC 1 taken on September 11, 2001 27
Figure 2-11. Stairwells in the WTC towers 29
Figure 2-12. Horizontal transfer floors in the WTC towers 30
Figure 2-13. Stairwell door signage as seen from inside stairwell A 30
Figure 2-14. Elevator riser diagram for WTC 1 and WTC 2 33
Figure 2-15. Fire command station in lobby of WTC on September 11, 2001, as seen from
mezzanine escalator, looking west 37
Figure 2-16. Elevator communication panel in the fire command station of WTC 1, as operated on
September 11,2001 40
Figure 4-1. Employment start date at WTC 69
Figure 5-1. WTC 1 impact, 8:46 a.m 73
Figure 5-2. Computer simulated impact damage to WTC 1 on floor 95 at 0.7 s after impact with
stairwells superimposed 74
Figure 5-3. Calculated damage to floors 93 through 98 in WTC 1 75
Figure 6-1. Observations of building damage after initial awareness but before beginning
evacuation in WTC 1 81
Figure 6-2. Fire Command Station in WTC 1 on September 11,2001 86
NIST NCSTAR 1-7, WTC Investigation ix
List of Figures
Figure 7-1. WTC 2 damage (computer simulated) at 0.62 s after impact with stairwells
superimposed 99
Figure 7-2. East-looking view through WTC 1 lobby as debris from WTC 2 impact travels past
lobby windows onto plaza 100
Figure 7-3. Calculated damage to floors 78 through 83 in WTC 2 (see NIST NCSTAR 1-6) 101
Figure 7-4. Observations of damage from tenant spaces in WTC 2 103
Figure 8-1. View from mezzanine level in WTC 1 looking east across WTC plaza, covered with
debris. 'Sphere,' artwork by Fritz Koening, can be seen on the plaza 109
Figure 8-2. Escalator from mezzanine to concourse level in WTC 1 on September 11, 2001 110
Figure 8-3. Evacuees leaving complex near WTC 5 1 10
Figure 8-4. Occupants seeking fresh air on floors 103 - 105 on north face of WTC 1 at 9:58:12
a.m 112
Figure 9-1. Collapse of WTC 2, viewed from north 116
Figure 9-2. Occupants leaving a previously stuck elevator on September 11, 2001, in the lobby of
WTC 1 123
Figure 10-1. Reported evacuation time for survivors of WTC 1 and WTC 2 126
Figure 10-2. Percentage of survivors remaining in the building for WTC 1 and WTC 2 126
Figure 10-3. Percentage of occupants remaining in the building for WTC 1 and WTC 2 127
Figure 10-4. Stairwell travel speed in WTC 1 for all stairwells 129
Figure 10-5. Model of evacuation delay 132
Figure 10-6. Causal model for predicting normalized stairwell evacuation time 136
Figure 10-7. Firefighter and occupants using a 44 in. stairwell in WTC 1 on September 11, 2001 154
Figure 10-8. Occupant and firefighter in bunker gear passing in a generic 44 in. stairwell 155
Figure 10-9. Simulation of full building evacuation of a WTC tower with different occupant loads. ... 163
NIST NCSTAR 1-7, WTC Investigation
List of Tables
Table P-l. Federal building and fire safety investigation of the WTC disaster xx
Table P-2. Public meetings and briefings of the WTC Investigation xxiii
Table 2-1. Use of floors in the WTC towers 12
Table 2-2. Plan view of stairwells in WTC 1 and WTC 2 20
Table 2-3. Minimum stairwell design for 42,850 ft 2 office plan 47
Table 3-1. Comparison of media interviewees and badge list 51
Table 3-2. Comparison of CNN victim list and badge list 52
Table 3-3. Disposition of the CATI sample and the total sample by tower 54
Table 3-4. Summary disposition rates by tower 55
Table 3-5. Response rate analysis for WTC 1 59
Table 3-6. Response rate analysis for WTC 2 59
Table 3-7. Example tabular face-to-face interview data entry 61
Table 4-1. Occupancy estimates on September 11, 2001, by tower 67
Table 4-2. Occupancy estimate uncertainty 68
Table 4-3. Age for telephone survey respondents in WTC 1 and WTC 2. a 68
Table 4-4. WTC fire drills in 12 months prior to September 11, 2001. a 70
Table 5-1. How survivors in WTC 1 became aware something was wrong on
September 11,2001 76
Table 5-2. Observations of building damage in WTC 1 when occupants first became aware
something was wrong on September 11, 2001 78
Table 6-1. Observations of conditions in WTC 1 before beginning evacuation 80
Table 6-2. Activities prior to evacuation reported in telephone survey by survivors from WTC 1 82
Table 6-3. Information received and sought prior to beginning evacuation in WTC 1 83
Table 6-4. Survivors perception of risk to self and others after airplane impact and prior to
entering stairwell in WTC 1 83
Table 6-5. Sources of help used by occupants prior to beginning their evacuation in WTC 1 84
Table 6-6. Single reason given by survivors to begin their evacuation in WTC 1 85
NISTNCSTAR 1-7, WTC Investigation
List of Tables
Table 6-7. Elapsed time (min) to initiate evacuation for survivors from WTC 1 88
Table 6-8. Stairwell chosen for evacuation in WTC 1 89
Table 6-9. How survivors in WTC 2 became aware that something was wrong on September 11,
2001 89
Table 6-10. Activities prior to evacuation reported in telephone surveys by survivors from
WTC 2 92
Table 6-11. Information received and sought prior to beginning evacuation in WTC 2 93
Table 6-12. Survivors' perception of risk upon airplane impact and prior to beginning their
evacuation in WTC 2 93
Table 6-13. How survivors received help prior to beginning their evacuation from WTC 2 94
Table 6-14. Single reason given by survivors to begin their evacuation from WTC 2 94
Table 6-15. Elapsed time (min) to initiate evacuation for survivors from WTC 2 95
Table 6-16. Stairwell chosen for evacuation in WTC 2 96
Table 8-1. Estimated times for survivors leaving the building in WTC 1 and WTC 2 105
Table 8-2. Location of WTC 2 survivors at or above floors of impact at 9:03 a.m 106
Table 8-3 . Observations of conditions during evacuation of WTC 1 and WTC 2 113
Table 8-4. Total evacuation time (min) for survivors from WTC 2 114
Table 9-1. Total evacuation time (min) for survivors from WTC 1 119
Table 9-2. World Trade Center decedent location 120
Table 10-1. Constraints to evacuation 151
Table 10-2. Egress simulation matrix 161
Table 10-3. Results for phased evacuation simulations 161
Table 10-4. Total building evacuation time (simulated) for various occupant loads 162
Table 10-5. Simulated evacuation results for fully-occupied WTC 1 and WTC 2 on
September 11,2001 164
xu
NIST NCSTAR 1-7, WTC Investigation
List of Acronyms and Abbreviations
Acronyms
ASTM ASTM International
BPS Building Performance Study
CATI Computer Assisted Telephone Interview
ESU Emergency Services Unit
FEMA Federal Emergency Management Agency
FDNY City of New York Fire Department
IBC International Building Code
NFPA National Fire Protection Association
NIST National Institute of Standards and Technology
NYC New York City
NYPD New York City Police Department
OSHA Occupational Safety and Health Administration
PANYNJ Port Authority of New York and New Jersey
PAPD Port Authority Police Department
WTC 1 World Trade Center 1 (North Tower)
WTC 2 World Trade Center 2 (South Tower)
WTC 7 World Trade Center 7
Abbreviations
°C
degrees Celsius
op
degrees Fahrenheit
ft
foot
ft 2
square foot
in.
inch
L
liter
m
meter
|am
micrometer
min
minute
s
second
NIST NCSTAR 1-7, WTC Investigation xiii
List of Acronyms and Abbreviations
This page intentionally left blank.
NIST NCSTAR 1-7, WTC Investigation
Metric Conversion Table
To convert from
to
Multiply by
AREA AND SECOND MOMENT OF AREA
square foot (ft )
square inch (in. )
square inch (in. )
square yard (yd )
square meter (m )
square meter (m )
square centimeter (cm )
square meter (m )
9.290 304 E-02
6.4516 E-04
6.4516 E+00
8.361 274 E-01
ENERGY (includes WORK)
kilowatt hour (kW ■ h)
quad (1015 BtuIT)
therm (U.S.)
ton of TNT (energy equivalent)
watt hour (W ■ h)
watt second (W • s)
joule (J)
joule (J)
joule (J)
joule (J)
joule (J)
joule (J)
3.6 E+06
1.055 056 E+18
1.054 804 E+08
4.184 E+09
3.6E+03
1.0 E+00
FORCE
dyne (dyn)
kilogram-force (kgf)
kilopond (kilogram-force) (kp)
kip (1 kip= 1,000 lbf)
kip (1 kip= 1,000 lbf)
pound-force (lbf)
newton (N)
newton (N)
newton (N)
newton (N)
kilonewton (kN)
newton (N)
1.0E-05
9.806 65 E+00
9.806 65 E+00
4.448 222 E+03
4.448 222 E+00
4.448 222 E+00
FORCE DIVIDED BY LENGTH
pound-force per foot (lbf/ft)
pound-force per inch (lbf/in.)
newton per meter (N/m)
newton per meter (N/m)
1.459 390 E+01
1.751 268E+02
HEAT FLOW RATE
calorieth per minute (calth/min)
calorieth per second (calth/s)
kilocalorieth per minute (kcalth/min)
kilocalorieth per second (kcalth/s)
watt (W)
watt (W)
watt (W)
watt (W)
6.973 333 E-02
4.184 E+00
6.973 333 E+01
4.184 E+03
NISTNCSTAR 1-7, WTC Investigation
Metric Conversion Table
To convert from
to
Multiply by
LENGTH
foot (ft)
inch (in)
inch (in.)
micron (m)
yard (yd)
meter (m)
meter (m)
centimeter (cm)
meter (m)
meter (m)
3.048 E-01
2.54 E-02
2.54 E+00
1.0E-06
9.144 E-01
MASS and MOMENT OF INERTIA
kilogram-force second squared
per meter (kgf ■ s 2 /m)
pound foot squared (lb ■ ft )
pound inch squared (lb ■ in. 2 )
ton, metric (t)
ton, short (2,000 lb)
kilogram (kg)
kilogram meter squared (kg ■ m )
kilogram meter squared (kg ■ m 2 )
kilogram (kg)
kilogram (kg)
9.806 65 E+00
4.214 011 E-02
2.926 397 E-04
1.0E+03
9.071 847 E+02
MASS DIVIDED BY AREA
pound per square foot (lb/ft )
pound per square inch
(not pound force) (lb/in. )
kilogram per square meter (kg/m ) 4.882 428 E+00
kilogram per square meter (kg/m ) 7.030 696 E+02
MASS DIVIDED BY LENGTH
pound per foot (lb/ft)
pound per inch (lb/in.)
pound per yard (lb/yd)
kilogram per meter (kg/m)
kilogram per meter (kg/m)
kilogram per meter (kg/m)
1.488 164 E+00
1.785 797 E+01
4.960 546 E-01
PRESSURE or STRESS (FORCE DIVIDED BY AREA)
kilogram-force per square centimeter (kgf/cm 2 ) pascal (Pa)
kilogram-force per square meter (kgf/m 2 )
pascal (Pa)
kilogram-force per square millimeter (kgf/mm ) pascal (Pa)
kip per square inch (ksi) (kip/in. ) pascal (Pa)
kip per square inch (ksi) (kip/in. ) kilopascal (kPa)
pound-force per square foot (lbf/ft ) pascal (Pa)
pound-force per square inch (psi) (lbf/in. ) pascal (Pa)
pound-force per square inch (psi) (lbf/in. 2 ) kilopascal (kPa)
psi (pound-force per square inch) (lbf/in. ) pascal (Pa)
psi (pound-force per square inch) (lbf/in. 2 ) kilopascal (kPa)
9.806 65 E+04
9.806 65 E+00
9.806 65 E+06
6.894 757 E+06
6.894 757 E+03
4.788 026 E+01
6.894 757 E+03
6.894 757 E+00
6.894 757 E+03
6.894 757 E+00
xvi
NIST NCSTAR 1-7, WTC Investigation
Metric Conversion Table
To convert from
to
Multiply by
TEMPERATURE
degree Celsius (°C)
degree centigrade
degree Fahrenheit (°F)
degree Fahrenheit (°F)
kelvin (K)
kelvin (K)
degree Celsius (°C)
degree Celsius (°C)
kelvin (K)
degree Celsius (°C)
T/K = t/°C + 273.15
t/°C » t /deg. cent.
t/°C = (t/°F-32)/1.8
T/K=(t/°F+459.67)/l.
t/°C = T/K 2 273.15
TEMPERATURE INTERVAL
degree Celsius (°C)
degree centigrade
degree Fahrenheit (°F)
degree Fahrenheit (°F)
degree Rankine (°R)
kelvin (K)
degree Celsius (°C)
degree Celsius (°C)
kelvin (K)
kelvin (K)
1.0E+00
1.0E+00
5.555 556 E-01
5.555 556 E-01
5.555 556 E-01
VELOCITY (includes SPEED)
foot per second (ft/s)
inch per second (in./s)
kilometer per hour (km/h)
mile per hour (mi/h)
mile per minute (mi/min)
meter per second (m/s)
meter per second (m/s)
meter per second (m/s)
kilometer per hour (km/h)
meter per second (m/s)
3.048 E-01
2.54 E-02
2.777 778 E-01
1.609 344 E+00
2.682 24 E+01
VOLUME (includes CAPACITY)
cubic foot (ft 3 )
cubic inch (in. )
cubic yard (yd 3 )
gallon (U.S.) (gal)
gallon (U.S.) (gal)
liter (L)
ounce (U.S. fluid) (fl oz)
ounce (U.S. fluid) (fl oz)
cubic meter (m )
cubic meter (m )
cubic meter (m 3 )
cubic meter (m 3 )
liter (L)
cubic meter (m 3 )
cubic meter (m )
milliliter (mL)
2.831 685 E-02
1.638 706 E-05
7.645 549 E-01
3.785 412 E-03
3.785 412 E+00
1.0 E-03
2.957 353 E-05
2.957 353 E+01
NISTNCSTAR 1-7, WTC Investigation
xvn
Metric Conversion Table
This page intentionally left blank.
xviii NIST NCSTAR 1-7, WTC Investigation
Preface
Genesis of This Investigation
Immediately following the terrorist attack on the World Trade Center (WTC) on September 11, 2001, the
Federal Emergency Management Agency (FEMA) and the American Society of Civil Engineers began
planning a building performance study of the disaster. The week of October 7, as soon as the rescue and
search efforts ceased, the Building Performance Study Team went to the site and began its assessment.
This was to be a brief effort, as the study team consisted of experts who largely volunteered their time
away from their other professional commitments. The Building Performance Study Team issued its
report in May 2002, fulfilling its goal "to determine probable failure mechanisms and to identify areas of
future investigation that could lead to practical measures for improving the damage resistance of buildings
against such unforeseen events."
On August 21, 2002, with funding from the U.S. Congress through FEMA, the National Institute of
Standards and Technology (NIST) announced its building and fire safety investigation of the WTC
disaster. On October 1, 2002, the National Construction Safety Team Act (Public Law 107-231), was
signed into law. The NIST WTC Investigation was conducted under the authority of the National
Construction Safety Team Act.
The goals of the investigation of the WTC disaster were:
• To investigate the building construction, the materials used, and the technical conditions that
contributed to the outcome of the WTC disaster.
• To serve as the basis for:
- Improvements in the way buildings are designed, constructed, maintained, and used;
- Improved tools and guidance for industry and safety officials;
- Recommended revisions to current codes, standards, and practices; and
- Improved public safety.
The specific objectives were:
1. Determine why and how WTC 1 and WTC 2 collapsed following the initial impacts of the
aircraft and why and how WTC 7 collapsed;
2. Determine why the injuries and fatalities were so high or low depending on location,
including all technical aspects of fire protection, occupant behavior, evacuation, and
emergency response;
3. Determine what procedures and practices were used in the design, construction, operation,
and maintenance of WTC 1, 2, and 7; and
4. Identify, as specifically as possible, areas in current building and fire codes, standards, and
practices that warrant revision.
NISTNCSTAR 1-7, WTC Investigation
Preface
NIST is a nonregulatory agency of the U.S. Department of Commerce's Technology Administration. The
purpose of NIST investigations is to improve the safety and structural integrity of buildings in the United
States, and the focus is on fact finding. NIST investigative teams are authorized to assess building
performance and emergency response and evacuation procedures in the wake of any building failure that
has resulted in substantial loss of life or that posed significant potential of substantial loss of life. NIST
does not have the statutory authority to make findings of fault nor negligence by individuals or
organizations. Further, no part of any report resulting from a NIST investigation into a building failure or
from an investigation under the National Construction Safety Team Act may be used in any suit or action
for damages arising out of any matter mentioned in such report (15 USC 281a, as amended by Public
Law 107-231).
Organization of the Investigation
The National Construction Safety Team for this Investigation, appointed by the then NIST Director,
Dr. Arden L. Bement, Jr., was led by Dr. S. Shyam Sunder. Dr. William L. Grosshandler served as
Associate Lead Investigator, Mr. Stephen A. Cauffman served as Program Manager for Administration,
and Mr. Harold E. Nelson served on the team as a private sector expert. The Investigation included eight
interdependent projects whose leaders comprised the remainder of the team. A detailed description of
each of these eight projects is available at http://wtc.nist.gov. The purpose of each project is summarized
in Table P-l, and the key interdependencies among the projects are illustrated in Fig. P-l.
Table P-l. Federal building and fire safety investigation of the WTC disaster.
Technical Area and Project Leader
Project Purpose
Analysis of Building and Fire Codes and
Practices; Project Leaders: Dr. H. S. Lew
and Mr. Richard W. Bukowski
Document and analyze the code provisions, procedures, and
practices used in the design, construction, operation, and
maintenance of the structural, passive fire protection, and
emergency access and evacuation systems of WTC 1, 2, and 7.
Baseline Structural Performance and
Aircraft Impact Damage Analysis; Project
Leader: Dr. Fahim H. Sadek
Analyze the baseline performance of WTC 1 and WTC 2 under
design, service, and abnormal loads, and aircraft impact damage on
the structural, fire protection, and egress systems.
Mechanical and Metallurgical Analysis of
Structural Steel; Project Leader: Dr. Frank
W. Gayle
Determine and analyze the mechanical and metallurgical properties
and quality of steel, weldments, and connections from steel
recovered from WTC 1, 2, and 7.
Investigation of Active Fire Protection
Systems; Project Leader: Dr. David
D. Evans; Dr. William Grosshandler
Investigate the performance of the active fire protection systems in
WTC 1 , 2, and 7 and their role in fire control, emergency response,
and fate of occupants and responders.
Reconstruction of Thermal and Tenability
Environment; Project Leader: Dr. Richard
G. Gann
Reconstruct the time-evolving temperature, thermal environment,
and smoke movement in WTC 1 , 2, and 7 for use in evaluating the
structural performance of the buildings and behavior and fate of
occupants and responders.
Structural Fire Response and Collapse
Analysis; Project Leaders: Dr. John
L. Gross and Dr. Therese P. McAllister
Analyze the response of the WTC towers to fires with and without
aircraft damage, the response of WTC 7 in fires, the performance
of composite steel-trussed floor systems, and determine the most
probable structural collapse sequence for WTC 1, 2, and 7.
Occupant Behavior, Egress, and Emergency
Communications; Project Leader: Mr. Jason
D. Averill
Analyze the behavior and fate of occupants and responders, both
those who survived and those who did not, and the performance of
the evacuation system.
Emergency Response Technologies and
Guidelines; Project Leader: Mr. J. Randall
Law son
Document the activities of the emergency responders from the time
of the terrorist attacks on WTC 1 and WTC 2 until the collapse of
WTC 7, including practices followed and technologies used.
NIST NCSTAR 1-7, WTC Investigation
Preface
WTC Building
Performance Study
Recommendations
Government,
Industry,
Professional,
Academic Inp
Public Inputs
Video/
Photographic
Records
Oral History Data
Emergency
Response
Records
Recovered
Structural Steel/
fMisr
NIST WTC Investigation Projects
, / Analysis of V^^
\ I Steel 1
■■n^^^/ Structural \
V Collapse J
/ Baseline
/ Performance
V & Impact
\ Damage .
/ [ Thermal and \
I Tenability J
\ Environment J
/ Analysis of \ .
I Codes and ) /
V Practices J 1
Active Fire \ V
Protection / 3- -\
^ ^ ( Emergency \
I Response /
1 Evacuation 1
Figure P-l. The eight projects in the federal building and fire safety
investigation of the WTC disaster.
National Construction Safety Team Advisory Committee
The NIST Director also established an advisory committee as mandated under the National Construction
Safety Team Act. The initial members of the committee were appointed following a public solicitation.
These were:
Paul Fitzgerald, Executive Vice President (retired) FM Global, National Construction Safety
Team Advisory Committee Chair
John Barsom, President, Barsom Consulting, Ltd.
John Bryan, Professor Emeritus, University of Maryland
David Collins, President, The Preview Group, Inc.
Glenn Corbett, Professor, John Jay College of Criminal Justice
Philip DiNenno, President, Hughes Associates, Inc.
NISTNCSTAR 1-7, WTC Investigation
xxi
•
•
•
Preface
Robert Hanson, Professor Emeritus, University of Michigan
Charles Thornton, Co-Chairman and Managing Principal, The Thornton-Tomasetti Group,
Inc.
Kathleen Tierney, Director, Natural Hazards Research and Applications Information Center,
University of Colorado at Boulder
Forman Williams, Director, Center for Energy Research, University of California at San
Diego
This National Construction Safety Team Advisory Committee provided technical advice during the
Investigation and commentary on drafts of the Investigation reports prior to their public release. NIST
has benefited from the work of many people in the preparation of these reports, including the National
Construction Safety Team Advisory Committee. The content of the reports and recommendations,
however, are solely the responsibility of NIST.
Public Outreach
During the course of this Investigation, NIST held public briefings and meetings (listed in Table P-2) to
solicit input from the public, present preliminary findings, and obtain comments on the direction and
progress of the Investigation from the public and the Advisory Committee.
NIST maintained a publicly accessible Web site during this Investigation at http://wtc.nist.gov. The site
contained extensive information on the background and progress of the Investigation.
NIST's WTC Public-Private Response Plan
The collapse of the WTC buildings has led to broad reexamination of how tall buildings are designed,
constructed, maintained, and used, especially with regard to major events such as fires, natural disasters,
and terrorist attacks. Reflecting the enhanced interest in effecting necessary change, NIST, with support
from Congress and the Administration, has put in place a program, the goal of which is to develop and
implement the standards, technology, and practices needed for cost-effective improvements to the safety
and security of buildings and building occupants, including evacuation, emergency response procedures,
and threat mitigation.
The strategy to meet this goal is a three-part NIST-led public-private response program that includes:
• A federal building and fire safety investigation to study the most probable factors that
contributed to post-aircraft impact collapse of the WTC towers and the 47-story WTC 7
building, and the associated evacuation and emergency response experience.
• A research and development (R&D) program to (a) facilitate the implementation of
recommendations resulting from the WTC Investigation, and (b) provide the technical basis
for cost-effective improvements to national building and fire codes, standards, and practices
that enhance the safety of buildings, their occupants, and emergency responders.
xxii NIST NCSTAR 1-7, WTC Investigation
Preface
Table P-2. Public meetings and briefings of the WTC Investigation.
Date
Location
Principal Agenda
June 24, 2002
New York City, NY
Public meeting: Public comments on the Draft Plan for the
pending WTC Investigation.
August 21, 2002
Gaithersburg, MD
Media briefing announcing the formal start of the Investigation.
December 9, 2002
Washington, DC
Media briefing on release of the Public Update and NIST request
for photographs and videos.
April 8, 2003
New York City, NY
Joint public forum with Columbia University on first-person
interviews.
April 29-30, 2003
Gaithersburg, MD
NCST Advisory Committee meeting on plan for and progress on
WTC Investigation with a public comment session.
May 7, 2003
New York City, NY
Media briefing on release of May 2003 Progress Report.
August 26-27, 2003
Gaithersburg, MD
NCST Advisory Committee meeting on status of the WTC
investigation with a public comment session.
September 17,2003
New York City, NY
Media and public briefing on initiation of first-person data
collection projects.
December 2-3, 2003
Gaithersburg, MD
NCST Advisory Committee meeting on status and initial results
and release of the Public Update with a public comment session.
February 12,2004
New York City, NY
Public meeting on progress and preliminary findings with public
comments on issues to be considered in formulating final
recommendations.
June 18,2004
New York City, NY
Media/public briefing on release of June 2004 Progress Report.
June 22-23, 2004
Gaithersburg, MD
NCST Advisory Committee meeting on the status of and
preliminary findings from the WTC Investigation with a public
comment session.
August 24, 2004
Northbrook, IL
Public viewing of standard fire resistance test of WTC floor
system at Underwriters Laboratories, Inc.
October 19-20, 2004
Gaithersburg, MD
NCST Advisory Committee meeting on status and near complete
set of preliminary findings with a public comment session.
November 22, 2004
Gaithersburg, MD
NCST Advisory Committee discussion on draft annual report to
Congress, a public comment session, and a closed session to
discuss pre-draft recommendations for WTC Investigation.
April 5, 2005
New York City, NY
Media and public briefing on release of the probable collapse
sequence for the WTC towers and draft reports for the projects on
codes and practices, evacuation, and emergency response.
June 23, 2005
New York City, NY
Media and public briefing on release of all draft reports for the
WTC towers and draft recommendations for public comment.
September 12-13,
2005
Gaithersburg, MD
NCST Advisory Committee meeting on disposition of public
comments and update to draft reports for the WTC towers.
September 13-15,
2005
Gaithersburg, MD
WTC Technical Conference for stakeholders and technical
community for dissemination of findings and recommendations
and opportunity for public to make technical comments.
• A dissemination and technical assistance program (DTAP) to (a) engage leaders of the
construction and building community in ensuring timely adoption and widespread use of
proposed changes to practices, standards, and codes resulting from the WTC Investigation
and the R&D program, and (b) provide practical guidance and tools to better prepare facility
owners, contractors, architects, engineers, emergency responders, and regulatory authorities
to respond to future disasters.
The desired outcomes are to make buildings, occupants, and first responders safer in future disaster
events.
NISTNCSTAR 1-7, WTC Investigation
xxm
Preface
National Construction Safety Team Reports on the WTC Investigation
A final report on the collapse of the WTC towers is being issued as NIST NCSTAR 1. A companion
report on the collapse of WTC 7 is being issued as NIST NCSTAR 1A. The present report is one of a set
that provides more detailed documentation of the Investigation findings and the means by which these
technical results were achieved. As such, it is part of the archival record of this Investigation. The titles
of the full set of Investigation publications are:
NIST (National Institute of Standards and Technology). 2005. Federal Building and Fire Safety
Investigation of the World Trade Center Disaster: Final Report on the Collapse of the World Trade
Center Towers. NIST NCSTAR 1. Gaithersburg, MD, September.
NIST (National Institute of Standards and Technology). 2006. Federal Building and Fire Safety
Investigation of the World Trade Center Disaster: Final Report on the Collapse of World Trade Center 7.
NIST NCSTAR 1A. Gaithersburg, MD.
Lew, H. S., R. W. Bukowski, and N. J. Carino. 2005. Federal Building and Fire Safety Investigation of
the World Trade Center Disaster: Design, Construction, and Maintenance of Structural and Life Safety
Systems. NIST NCSTAR 1-1. National Institute of Standards and Technology. Gaithersburg, MD,
September.
Fanella, D. A., A. T. Derecho, and S. K. Ghosh. 2005. Federal Building and Fire Safety
Investigation of the World Trade Center Disaster: Design and Construction of Structural Systems.
NIST NCSTAR 1-1 A. National Institute of Standards and Technology. Gaithersburg, MD,
September.
Ghosh, S. K., and X. Liang. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Comparison of Building Code Structural Requirements. NIST
NCSTAR 1-1B. National Institute of Standards and Technology. Gaithersburg, MD, September.
Fanella, D. A., A. T. Derecho, and S. K. Ghosh. 2005. Federal Building and Fire Safety
Investigation of the World Trade Center Disaster: Maintenance and Modifications to Structural
Systems. NIST NCSTAR 1-lC. National Institute of Standards and Technology. Gaithersburg,
MD, September.
Grill, R. A., and D. A. Johnson. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Fire Protection and Life Safety Provisions Applied to the Design and
Construction of World Trade Center 1, 2, and 7 and Post-Construction Provisions Applied after
Occupancy. NIST NCSTAR 1-1D. National Institute of Standards and Technology. Gaithersburg,
MD, September.
Razza, J. C, and R. A. Grill. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Comparison of Codes, Standards, and Practices in Use at the Time of the
Design and Construction of World Trade Center 1, 2, and 7. NIST NCSTAR 1-1E. National
Institute of Standards and Technology. Gaithersburg, MD, September.
Grill, R. A., D. A. Johnson, and D. A. Fanella. 2005. Federal Building and Fire Safety
Investigation of the World Trade Center Disaster: Comparison of the 1968 and Current (2003) New
xxiv NIST NCSTAR 1-7, WTC Investigation
Preface
York City Building Code Provisions. NIST NCSTAR 1-1F. National Institute of Standards and
Technology. Gaithersburg, MD, September.
Grill, R. A., and D. A. Johnson. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Amendments to the Fire Protection and Life Safety Provisions of the New
York City Building Code by Local Laws Adopted While World Trade Center 1, 2, and 7 Were in
Use. NIST NCSTAR 1-1G. National Institute of Standards and Technology. Gaithersburg, MD,
September.
Grill, R. A., and D. A. Johnson. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Post-Construction Modifications to Fire Protection and Life Safety Systems
of World Trade Center 1 and 2. NIST NCSTAR 1-1H. National Institute of Standards and
Technology. Gaithersburg, MD, September.
Grill, R. A., D. A. Johnson, and D. A. Fanella. 2005. Federal Building and Fire Safety Investigation
of the World Trade Center Disaster: Post-Construction Modifications to Fire Protection, Life
Safety, and Structural Systems of World Trade Center 7. NIST NCSTAR 1-1 1. National Institute of
Standards and Technology. Gaithersburg, MD, September.
Grill, R. A., and D. A. Johnson. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Design, Installation, and Operation of Fuel System for Emergency Power in
World Trade Center 7. NIST NCSTAR 1-1 J. National Institute of Standards and Technology.
Gaithersburg, MD, September.
Sadek, F. 2005. Federal Building and Fire Safety Investigation of the World Trade Center Disaster:
Baseline Structural Performance and Aircraft Impact Damage Analysis of the World Trade Center
Towers. NIST NCSTAR 1-2. National Institute of Standards and Technology. Gaithersburg, MD,
September.
Faschan, W. J., and R. B. Garlock. 2005. Federal Building and Fire Safety Investigation of the
World Trade Center Disaster: Reference Structural Models and Baseline Performance Analysis of
the World Trade Center Towers. NIST NCSTAR 1-2A. National Institute of Standards and
Technology. Gaithersburg, MD, September.
Kirkpatrick, S. W., R. T. Bocchieri, F. Sadek, R. A. MacNeill, S. Holmes, B. D. Peterson,
R. W. Cilke, C. Navarro. 2005. Federal Building and Fire Safety Investigation of the World Trade
Center Disaster: Analysis of Aircraft Impacts into the World Trade Center Towers, NIST
NCSTAR 1-2B. National Institute of Standards and Technology. Gaithersburg, MD, September.
Gayle, F. W., R. J. Fields, W. E. Luecke, S. W. Banovic, T. Foecke, C. N. McCowan, T. A. Siewert, and
J. D. McColskey. 2005. Federal Building and Fire Safety Investigation of the World Trade Center
Disaster: Mechanical and Metallurgical Analysis of Structural Steel. NIST NCSTAR 1-3. National
Institute of Standards and Technology. Gaithersburg, MD, September.
Luecke, W. E., T. A. Siewert, and F. W. Gayle. 2005. Federal Building and Fire Safety
Investigation of the World Trade Center Disaster: Contemporaneous Structural Steel
Specifications. NIST Special Publication 1-3 A. National Institute of Standards and Technology.
Gaithersburg, MD, September.
NIST NCSTAR 1-7, WTC Investigation
Preface
Banovic, S. W. 2005. Federal Building and Fire Safety Investigation of the World Trade Center
Disaster: Steel Inventory and Identification. NIST NCSTAR 1-3B. National Institute of Standards
and Technology. Gaithersburg, MD, September.
Banovic, S. W., and T. Foecke. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Damage and Failure Modes of Structural Steel Components. NIST
NCSTAR 1-3C. National Institute of Standards and Technology. Gaithersburg, MD, September.
Luecke, W. E., J. D. McColskey, C. N. McCowan, S. W. Banovic, R. J. Fields, T. Foecke,
T. A. Siewert, and F. W. Gayle. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Mechanical Properties of Structural Steels. NIST NCSTAR 1-3D.
National Institute of Standards and Technology. Gaithersburg, MD, September.
Banovic, S. W., C. N. McCowan, and W. E. Luecke. 2005. Federal Building and Fire Safety
Investigation of the World Trade Center Disaster: Physical Properties of Structural Steels. NIST
NCSTAR 1-3E. National Institute of Standards and Technology. Gaithersburg, MD, September.
Evans, D. D., R. D. Peacock, E. D. Kuligowski, W. S. Dols, and W. L. Grosshandler. 2005. Federal
Building and Fire Safety Investigation of the World Trade Center Disaster: Active Fire Protection
Systems. NIST NCSTAR 1-4. National Institute of Standards and Technology. Gaithersburg, MD,
September.
Kuligowski, E. D., D. D. Evans, and R. D. Peacock. 2005. Federal Building and Fire Safety
Investigation of the World Trade Center Disaster: Post-Construction Fires Prior to September 11,
2001. NIST NCSTAR 1-4A. National Institute of Standards and Technology. Gaithersburg, MD,
September.
Hopkins, M., J. Schoenrock, and E. Budnick. 2005. Federal Building and Fire Safety Investigation
of the World Trade Center Disaster: Fire Suppression Systems. NIST NCSTAR 1-4B. National
Institute of Standards and Technology. Gaithersburg, MD, September.
Keough, R. J., and R. A. Grill. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Fire Alarm Systems. NIST NCSTAR 1-4C. National Institute of Standards
and Technology. Gaithersburg, MD, September.
Ferreira, M. J., and S. M. Strege. 2005. Federal Building and Fire Safety Investigation of the
World Trade Center Disaster: Smoke Management Systems. NIST NCSTAR 1-4D. National
Institute of Standards and Technology. Gaithersburg, MD, September.
Gann, R. G., A. Hamins, K. B. McGrattan, G. W. Mulholland, H. E. Nelson, T. J. Ohlemiller,
W. M. Pitts, and K. R. Prasad. 2005. Federal Building and Fire Safety Investigation of the World Trade
Center Disaster: Reconstruction of the Fires in the World Trade Center Towers. NIST NCSTAR 1-5.
National Institute of Standards and Technology. Gaithersburg, MD, September.
Pitts, W. M., K. M. Butler, and V. Junker. 2005. Federal Building and Fire Safety Investigation of
the World Trade Center Disaster: Visual Evidence, Damage Estimates, and Timeline Analysis.
NIST NCSTAR 1-5 A. National Institute of Standards and Technology. Gaithersburg, MD,
September.
xxvi NIST NCSTAR 1-7, WTC Investigation
Preface
Hamins, A., A. Maranghides, K. B. McGrattan, E. Johnsson, T. J. Ohlemiller, M. Donnelly,
J. Yang, G. Mulholland, K. R. Prasad, S. Kukuck, R. Anleitner and T. McAllister. 2005. Federal
Building and Fire Safety Investigation of the World Trade Center Disaster: Experiments and
Modeling of Structural Steel Elements Exposed to Fire. NIST NCSTAR 1-5B. National Institute of
Standards and Technology. Gaithersburg, MD, September.
Ohlemiller, T. J., G. W. Mulholland, A. Maranghides, J. J. Filliben, and R. G. Gann. 2005. Federal
Building and Fire Safety Investigation of the World Trade Center Disaster: Fire Tests of Single
Office Workstations. NIST NCSTAR 1-5C. National Institute of Standards and Technology.
Gaithersburg, MD, September.
Gann, R. G., M. A. Riley, J. M. Repp, A. S. Whittaker, A. M. Reinhorn, and P. A. Hough. 2005.
Federal Building and Fire Safety Investigation of the World Trade Center Disaster: Reaction of
Ceiling Tile Systems to Shocks. NIST NCSTAR 1-5D. National Institute of Standards and
Technology. Gaithersburg, MD, September.
Hamins, A., A. Maranghides, K. B. McGrattan, T. J. Ohlemiller, and R. Anleitner. 2005. Federal
Building and Fire Safety Investigation of the World Trade Center Disaster: Experiments and
Modeling of Multiple Workstations Burning in a Compartment. NIST NCSTAR 1-5E. National
Institute of Standards and Technology. Gaithersburg, MD, September.
McGrattan, K. B., C. Bouldin, and G. Forney. 2005. Federal Building and Fire Safety
Investigation of the World Trade Center Disaster: Computer Simulation of the Fires in the World
Trade Center Towers. NIST NCSTAR 1-5F. National Institute of Standards and Technology.
Gaithersburg, MD, September.
Prasad, K. R., and H. R. Baum. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Fire Structure Interface and Thermal Response of the World Trade Center
Towers. NIST NCSTAR 1-5G. National Institute of Standards and Technology. Gaithersburg,
MD, September.
Gross, J. L., and T. McAllister. 2005. Federal Building and Fire Safety Investigation of the World Trade
Center Disaster: Structural Fire Response and Probable Collapse Sequence of the World Trade Center
Towers. NIST NCSTAR 1-6. National Institute of Standards and Technology. Gaithersburg, MD,
September.
Carino, N. J., M. A. Starnes, J. L. Gross, J. C. Yang, S. Kukuck, K. R. Prasad, and R. W. Bukowski.
2005. Federal Building and Fire Safety Investigation of the World Trade Center Disaster: Passive
Fire Protection. NIST NCSTAR 1-6A. National Institute of Standards and Technology.
Gaithersburg, MD, September.
Gross, J., F. Hervey, M. Izydorek, J. Mammoser, and J. Treadway. 2005. Federal Building and
Fire Safety Investigation of the World Trade Center Disaster: Fire Resistance Tests of Floor Truss
Systems. NIST NCSTAR 1-6B. National Institute of Standards and Technology. Gaithersburg,
MD, September.
Zarghamee, M. S., S. Bolourchi, D. W. Eggers, O. O. Erbay, F. W. Kan, Y. Kitane, A. A. Liepins,
M. Mudlock, W. I. Naguib, R. P. Ojdrovic, A. T. Sarawit, P. R Barrett, J. L. Gross, and
NIST NCSTAR 1-7, WTC Investigation xxvii
Preface
T. P. McAllister. 2005. Federal Building and Fire Safety Investigation of the World Trade Center
Disaster: Component, Connection, and Subsystem Structural Analysis. NIST NCSTAR 1-6C.
National Institute of Standards and Technology. Gaithersburg, MD, September.
Zarghamee, M. S., Y. Kitane, O. O. Erbay, T. P. McAllister, and J. L. Gross. 2005. Federal
Building and Fire Safety Investigation of the World Trade Center Disaster: Global Structural
Analysis of the Response of the World Trade Center Towers to Impact Damage and Fire. NIST
NCSTAR 1-6D. National Institute of Standards and Technology. Gaithersburg, MD, September.
McAllister, T., R. W. Bukowski, R. G. Gann, J. L. Gross, K. B. McGrattan, H. E. Nelson, L. Phan,
W. M. Pitts, K. R. Prasad, F. Sadek. 2006. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Structural Fire Response and Probable Collapse Sequence of World Trade
Center 7. (Provisional). NIST NCSTAR 1-6E. National Institute of Standards and Technology.
Gaithersburg, MD.
Gilsanz, R., V. Arbitrio, C. Anders, D. Chlebus, K. Ezzeldin, W. Guo, P. Moloney, A. Montalva,
J. Oh, K. Rubenacker. 2006. Federal Building and Fire Safety Investigation of the World Trade
Center Disaster: Structural Analysis of the Response of World Trade Center 7 to Debris Damage
and Fire. (Provisional). NIST NCSTAR 1-6F. National Institute of Standards and Technology.
Gaithersburg, MD.
Kim, W. 2006. Federal Building and Fire Safety Investigation of the World Trade Center
Disaster: Analysis of September 11, 2001, Seismogram Data. (Provisional). NIST NCSTAR 1-6G.
National Institute of Standards and Technology. Gaithersburg, MD.
Nelson, K. 2006. Federal Building and Fire Safety Investigation of the World Trade Center
Disaster: The Con Ed Substation in World Trade Center 7. (Provisional). NIST NCSTAR 1-6H.
National Institute of Standards and Technology. Gaithersburg, MD.
Averill, J. D., D. S. Mileti, R. D. Peacock, E. D. Kuligowski, N. Groner, G. Proulx, P. A. Reneke, and
H. E. Nelson. 2005. Federal Building and Fire Safety Investigation of the World Trade Center Disaster:
Occupant Behavior, Egress, and Emergency Communication. NIST NCSTAR 1-7. National Institute of
Standards and Technology. Gaithersburg, MD, September.
Fahy, R., and G. Proulx. 2005. Federal Building and Fire Safety Investigation of the World Trade
Center Disaster: Analysis of Published Accounts of the World Trade Center Evacuation. NIST
NCSTAR 1-7A. National Institute of Standards and Technology. Gaithersburg, MD, September.
Zmud, J. 2005. Federal Building and Fire Safety Investigation of the World Trade Center
Disaster: Technical Documentation for Survey Administration. NIST NCSTAR 1-7B. National
Institute of Standards and Technology. Gaithersburg, MD, September.
Lawson, J. R., and R. L. Vettori. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: The Emergency Response Operations. NIST NCSTAR 1-8. National Institute of
Standards and Technology. Gaithersburg, MD, September.
xxviii NIST NCSTAR 1-7, WTC Investigation
Acknowledgments
The authors wish to acknowledge the contributions of a number of people, without whom, this project
would not have been successful.
Robyn Gershon, Erin Hogan, and Stephen Morse, Columbia University; Nora Marshall and others from
the National Transportation Safety Board; Henry Quarantelli, and Benigno Aguirre from the Disaster
Research Center, University of Delaware; Linda Bourque - University of California - Los Angeles;
Richard Mendelson, Occupational Safety and Health Administration; Jake Pauls; and Edwina Juillet all
provided professional insight into issues, procedures, and pitfalls of conducting research involving human
behavior in disasters.
Rita Fahy, National Fire Protection Association analyzed media accounts of the September 11, 2001,
attacks and provided the EXIT89 model and support in its use.
Jay Cohen and George Duke, employees of the Corporation Counsel of the City of New York, provided
access to emergency communications and 9-1-1 records that provided details of the experiences and
conditions above the floors of impact in the WTC towers.
Frank Lombardi, Saroj Bhol, Alan Reiss, and Nancy Seliga, PANYNJ provided access to numerous Port
Authority records on the World Trade Center egress system and building procedures and facilitated access
to Port Authority personnel with extensive knowledge of the building operation and the events of
September 11,2001.
Patti Adler, Julie Gailus, Janet Jacobs, and Lori Peek - University of Colorado - Boulder; Martha van
Haitsma and Virginia Bartot - University of Chicago provided guidance on the analysis of qualitative
interview data.
Peter Thompson, Integrated Environmental Solutions, Inc.; Ed Galea and Steven Gwynne, University of
Greenwich - UK; and Jeremy Fraser-Mitchell, BRE - FRS, UK provided support and suggestions on
appropriate techniques for modeling evacuation in high-rise structures.
Finally, National Institute of Standards and Technology would like to gratefully acknowledge the more
than 1,000 survivors and families of victims of the September 11, 2001, attacks that generously agreed to
be interviewed about their difficult and often tragic experiences on September 11, 2001. Their detailed
accounts provide a continuing memory of the attacks and the victims of September 11, 2001.
NIST NCSTAR 1-7, WTC Investigation xxix
Acknowledgments
This page intentionally left blank.
NIST NCSTAR 1-7, WTC Investigation
Executive Summary
e.i OVERVIEW
While most attention has properly focused on the nearly three thousand people who lost their lives at the
World Trade Center (WTC) site on September 11, 2001, five times that many people successfully
evacuated from the WTC towers due to heroic efforts of occupants, as well as emergency responders.
Understanding why many, yet not all, survived the WTC attacks was one of the four objectives of the
Federal building and fire safety investigation of the WTC disaster led by the National Institute of
Standards and Technology (NIST).
Success in evacuating a building in an emergency can be characterized by two quantities: the time people
needed to evacuate and the time available for them to do so. To the extent the first time exceeded the
second, it follows that there will be casualties. When the second time exceeds the first, perhaps by some
suitable margin, nearly all should be able to evacuate the building.
For the WTC towers, the times available for escape were cataclysmically established by the collapses of
the buildings. Those times were not known in advance by the building occupants or the responders. The
times were also considerably shorter, by a factor of three or four, than the time needed to clear the tenant
spaces of WTC 1 following the 1993 bombing and an additional factor of two shorter than the time
needed to clear the last person from the elevators in the building. Further, some occupants would have
been unable to evacuate the buildings given any amount of time due to injuries, entrapment, and/or toxic
exposure.
NIST examined the design of the building, the behavior of the people, and the evacuation process in detail
to ascertain the factors that factored prominently in the time needed for evacuation.
In order to accomplish this objective, numerous sources of data were collected and analyzed, including:
over 1,000 new interviews with survivors; a collection of over 700 published interviews with
WTC survivors; 9-1-1 emergency calls; transcripts of emergency communication among building
personnel and emergency responders; historical building design drawings, memoranda, and calculations;
building modifications and upgrades; formal complaints filed with Occupational Safety and Health
Administration; and other relevant material.
There were three forms of interviews with survivors: 803 telephone interviews, over 225 face-to-face
interviews, and 6 focus groups. The telephone interviewees were randomly selected using independent
proportionate stratification from a list of occupants who had badges to enter WTC 1 or WTC 2 on
September 1 1 , 200 1 . In other words, each occupant of a particular tower had an equal probability of
being selected. Roughly 400 occupants in each tower were interviewed in order to achieve a high level of
statistical precision within each tower. Reported percentages from tower-specific survey data (n=400)
exhibited sampling errors no greater than 2.5 percentage points, and 95 percent confidence intervals of
percentages are no greater than ± 5 percentage points. This level of precision was more than adequate for
examining characteristics of occupants and egress attributes. With telephone interview results, primary
statistical analyses were in the form of tabulations and linear statistics (e.g., reporting of percentages and
NISTNCSTAR 1-7, WTC Investigation
Executive Summary
average/means). The telephone interview results enabled a scientific projection of the population and
distribution of occupants in WTC 1 and WTC 2, as well as causal modeling and multivariate regression
analysis to explore fundamental egress issues such as sources of evacuation delay.
The objective of the face-to-face interviews was to gather first-hand accounts and observations of the
activities and events inside the buildings on the morning of September 11. This approach identified
unknown information, aided in the evaluation of technical hypotheses, and explored motivations for
occupant behaviors, while allowing for comparisons to the telephone interview data. There was no
recording of the face-to-face interviews, other than random selections, with consent of respondents, for
quality control purposes. A typical face-to-face interview averaged approximately two hours. The
methodology for the face-to-face interviews was a synthesis of two established methodologies, designed
to assist survivors in providing comprehensive and accurate accounts of their evacuation, given the
latency between experience and interview. Some groups of occupants were specifically sought in order to
explore targeted unknowns. These included occupants near the floors of impact, witnesses to fireballs,
mobility-impaired occupants, floor wardens, building personnel with emergency response responsibilities,
family members who spoke to an occupant after 8:46:30 a.m., and occupants from regions of the building
not addressed by other groups in order to ensure adequate interview coverage for all areas of both towers.
Six focus groups were conducted in order to elicit accurate group representations of specific events or
themes and complement the findings of the telephone and face-to-face interviews. The focus groups and
the corresponding objectives were:
1. Occupants located near the floors of impact: to explore the extent of the building damage and
how the damage influenced the evacuation process.
2. Floor wardens: to explore the implementation of the floor warden procedures and the effect
those actions had on the evacuation of the occupants on a floor and the evacuation of the floor
warden.
3. Mobility-impaired occupants: to explore the effect of a disability on the evacuation of the
occupant and any other individuals who may have assisted or otherwise been affected by the
evacuee.
4. Persons with building responsibilities: to capture the unique perspective of custodians,
security, maintenance, or other building staff.
5. Randomly selected evacuees in WTC 1 : to further explore the variables from the causal
modeling which best explained evacuation delay and normalized stairwell evacuation time,
including environmental cues, floor, and activities.
6. Randomly selected evacuees in WTC 2: to further explore variables used in the causal
modeling that best explained evacuation delay, including environmental cues, floor, risk
perception, and use of elevators.
NIST documented the WTC egress system, including the location of the three primary stairwells, exit
doors, core hallways, transfer corridors, wall construction, location and layout of the 100+ elevators in
each tower, and emergency communication devices. The design of the egress system was compared to
xxxii NIST NCSTAR 1-7, WTC Investigation
Executive Summary
building code requirements of the New York City Building Code, National Fire Protection
Association 101 (Life Safety Code), and International Building Code.
NIST documented the emergency procedures, both as they were planned to be carried out, as well as how
they were actually implemented on September 11, 2001. The procedures included responsibilities for
tenant safety through the floor warden system; pre-planned content of public address system
announcements (which varied from public address system announcements made on September 11, 2001);
responsibilities of the fire safety director, deputy fire safety director, building security, and supervisors of
various contractors (including mechanical, vertical transportation, and electrical). Additionally,
interaction among responding agencies such as the Port Authority of New York and New Jersey, the Port
Authority Police Department, the New York City Police Department (NYPD), the New York City Fire
Department, and contract security were documented.
NIST estimates that there were 8,900 ± 750 people in WTC 1 at 8:46:30 a.m. on September 11, 2001.
Similarly, NIST estimates that there were 8,540 ± 920 people inside WTC 2 at 8:46:30 a.m. New York
City officially announced 2,749 fatalities at the WTC complex, including emergency responders, airplane
passengers and crew (but not hijackers), and bystanders. NIST estimated that of the 17,400 ± 1,180
occupants inside WTC 1 and WTC 2 at 8:46:30 a.m., 2,146 to 2,163 perished. No information could be
found for 17 persons. More than twice as many occupants were killed in WTC 1 as WTC 2, largely due
to the fact that occupants in WTC 2 used the 16 minutes between the attacks on WTC 1 and WTC 2 to
begin evacuating, including the use of elevators by some occupants in WTC 2.
The demographic characteristics of the evacuees was explored where the characteristics were relevant to
the evacuation on September 11, 2001. Few differences in the characteristics of WTC 1 or WTC 2 were
observed. Men outnumbered women roughly two to one. The average age was mid-forties. The mean
length of employment at the WTC site was almost 6 years, while the median was 2 and 3 years for
WTC 1 and 2, respectively. Sixteen percent of 2001 WTC evacuees were also present during the 1993
bombing, although many other occupants were also knowledgeable about the 1993 evacuation. Two-
thirds of the occupants had participated in at least one fire drill during the 12 months immediately prior to
September 11, 2001. Eighteen percent did not recall whether they had participated in a fire drill during
that time period and 18 percent reported that they did not participate in a fire drill during that time period.
In WTC 1 , all three stairwells and the elevators were destroyed in the impact region, extending as low as
floor 92. No occupant evacuated from above the 91st floor, although some survived until the building
collapsed after 102 minutes. Helicopter rescue from the roof was considered by an NYPD aviation unit,
but deemed not possible due to the heat and smoke from the building fire. Occupants of both towers
delayed initiating their evacuation after WTC 1 was hit. In WTC 1 , the median time to initiate evacuation
was 3 minutes for occupants from the ground floor to floor 76, and 5 minutes for occupants near the
impact region (floors 77 to 91). Occupants observed various types of impact indicators throughout the
building, including wall, partition, and ceiling damage and fire and smoke conditions. The most severe
damage was observed near the impact region, fatally trapping some occupants. Announcements in
WTC 1 were not heard by the occupants, despite repeated attempts from the lobby fire command station
to order an evacuation. Damage to critical communications hardware prevented announcement
transmission. Evacuation rates reached a peak, steady-state in approximately 5 minutes, and remained
roughly constant until the collapse of WTC 2, when the rate in WTC 1 slowed to about one-fifth of the
NIST NCSTAR 1-7, WTC Investigation xxxiii
Executive Summary
peak, steady-state. WTC 1 collapsed at 10:28:22 a.m., resulting in approximately 1,500 occupant deaths,
107 of which were estimated to be below the 92nd floor.
The evacuation of WTC 2 was markedly different from the evacuation of WTC 1 . There was a 1 6 minute
period after WTC 1 was attacked, but before WTC 2 was attacked. During this time period, occupants
were forced to decide whether to remain inside WTC 2, and if they decided to leave, they had to choose
between using one of the three stairwells or using an elevator. Further complicating this decision process
were multiple, conflicting announcements around 9:00 a.m., first instructing occupants to return to their
offices, and then within one minute of impact, instructing them to begin an evacuation if conditions on
their floor warranted that decision. Over 90 percent of WTC 2 survivors started to evacuate the building
prior to its being attacked. Sixteen percent of the survivors used elevators to evacuate. Approximately
75 percent of the occupants who were above the 78th floor (the lowest floor of impact) descended to at
least below the impact region prior to the attack on WTC 2. Over 40 percent of the survivors had left
WTC 2 prior to 9:02:59 a.m. After WTC 2 was attacked, at least 18 individuals used Stairwell A, located
in the northwest corner and furthest from the impact damage, to descend below the 78th floor to evacuate
the building. Additional public address announcements were made after the airplane strike on WTC 2,
although occupants who survived generally did not hear those announcements. After the initial peak in
evacuation rate, the rate reached a steady-state similar to the rate observed in WTC 1 until approximately
20 minutes prior to collapse of WTC 2. The evacuation rate during the final 20 minutes dropped
significantly, likely due to a decreased number of occupants remaining in the egress system below the
78th floor. NIST analysis indicated only 1 1 occupants initially below the 78th floor were killed when
WTC 2 collapsed at 9:58:59 a.m. Overall, NIST estimated that 630 occupants of WTC 2 perished.
Using the statistical power of the telephone interview results, causal models were constructed to explain
both evacuation initiation delay and average stairwell travel time per floor. The factors that best predicted
evacuation initiation delay in WTC 1 were (1) which floor the respondent was on when WTC 1 was
attacked, (2) whether occupants encountered environmental cues, and (3) seeking additional information
(or milling) about the nature of the event. In WTC 2, the same process occurred as in WTC 1, except that
perceived risk (sense of immediate danger) was a predictor of seeking additional information (along with
floor and environmental cues). Analyses explored factors that affected time spent in the stairwells in
WTC 1 exiting the building. The floor an occupant was on when WTC 1 was attacked (distance to safety)
increased the probability of encountering an environmental cue (smoke, damage, fire, etc). Additionally,
being on a higher floor predicted greater evacuation initiation delay times and encountering
environmental cues, which predicted higher normalized stairwell travel time. Independently, interrupting
evacuation for any reason increased the normalized stairwell travel time.
Constraints or aids to the evacuation progress were documented. Building announcements were cited by
many in WTC 2 as a constraint to their evacuation, principally due to the 9:00 a.m. announcement
instructing occupants to return to their work spaces. Crowdedness in the stairwells, firefighter
counterflow, lack of instructions and information, as well as injured or disabled evacuees in the stairwells
were the most frequently reported obstacles to evacuation. The most commonly mentioned forms of aid
were assistance from coworkers and emergency responders and the photoluminescent markings in
stairwells. Six percent of survivors in WTC 1 and WTC 2 reported a mobility impairment which slowed
their evacuation. Sometimes the evacuation speed of others in the immediate area slowed down occupant
evacuation speed. Recent pre-existing injuries, medications, or medical treatments were the most
commonly reported mobility impairments, while a small number used wheelchairs, were pregnant, or
xxxiv NIST NCSTAR 1-7, WTC Investigation
Executive Summary
were elderly. A rest station for mobility-impaired occupants was established in WTC 1 somewhere
between floors 12 and 20. Less than 10 minutes prior to the collapse of WTC 1, the occupants and
helpers on the floor were ordered to evacuate, although it remains unclear whether all rest station
residents survived.
Minutes prior to the collapse of WTC 2, an NYPD Emergency Services Unit (ESU) officer radioed from a
floor in the 20s to the outside that he was having trouble ascending the stairwell due to the large number
of occupants descending (Interview 24 NYPD [NIST 2004]). While the origin of the occupants remains
unknown, only 1 1 occupants who started evacuating below the impact region were known not to have
survived.
Multiple evacuation models were used to simulate different WTC tower evacuations, subject to a number
of assumptions. The goal of the modeling was to frame an understanding of actual evacuation findings on
September 11, 2001. Simulations demonstrated that a phased evacuation (also known as defend-in-place,
whereupon occupants on the fire floor and the immediately surrounding floors descend to three floors
below the fire floor) would have taken between 4 minutes to complete (without delays in evacuation
initiation) and 11 minutes to complete (with evacuation initiation delays between and 10 minutes).
Total evacuation of a tower assuming a full occupant load without visitors (19,800) would have required
as few as 92 minutes to 1 12 minutes. With visitors (total population 25,500 people) total evacuation
would have required as little as 1 14 minutes to 142 minutes. The ranges reflect two different model
outputs, each assuming two different delay times (no delay and a 10 minute distribution of delay times).
An evacuation simulation for 8,800 people (approximately the number present in each tower on
September 11, 2001) in the absence of any damage to the building, would have required at least
52 minutes to 71 minutes, depending on the model or the delay times. Finally, the EXODUS model was
'calibrated' to approximate the gross evacuation rates observed in WTC 1 and WTC 2 on
September 11, 2001. Once the model input necessary to approximate the observables was determined,
additional occupants were added in order to estimate how many occupants might have been unable to
evacuate on September 11, 2001 (given the damage to the building and observed delay times) if the
buildings had had larger occupant loads. NIST estimated that approximately 14,000 occupants would
have been unable to evacuate from WTC 1 and WTC 2 on September 11, 2001, had the starting building
population been 19,800 in each building.
E.2 REFERENCE
NIST (National Institute of Standards and Technology). 2004. NIST WTC Emergency Responder
Interview Data Set. Gaithersburg, MD.
NISTNCSTAR 1-7, WTC Investigation
Executive Summary
This page intentionally left blank.
xxxvi NIST NCSTAR 1-7, WTC Investigation
Chapter 1
Background and Introduction
On the morning of September 11, 2001, the World Trade Center (WTC) in New York City was attacked
by hijacked commercial airplanes. The collision with each tower (WTC 1 at 8:46:30 a.m. and WTC 2 at
9:02:59 a.m.) produced significant structural damage. The impact generated a large, luminous external
fireball that consumed a portion of the jet fuel, with the remaining fuel acting as an ignition source for the
combustible material within each tower. At 9:58:59 a.m., 56 minutes after it was struck, WTC 2
collapsed due to a combination of the aircraft impact damage and subsequent fire. WTC 1 stood until
10:28:22 a.m.
This report provides an analysis of the overall evacuation of WTC 1 and WTC 2. The two towers were
nearly identical buildings in height, geometry, and architectural features. The evacuation processes in
these two buildings displayed both distinct similarities and differences. This report also focuses on the
behaviors of the occupants, actions of the building personnel and emergency responders (covered more
completely in NIST NCSTAR 1-8 1 ), and the interactions among all three. This report documents the
performance of the emergency egress system.
This chapter begins with a discussion of significant egress events. It then reviews the design of the
WTC egress system and emergency procedures, and outlines the National Institute of Standards and
Technology (NIST) data collection methodology. The events of September 11, 2001, are detailed as they
relate to the evacuation, including the moments prior to 8:46:30 a.m. Finally, an analysis of the events
establishes key evacuation findings and conclusions. Note that individuals shown in photographs in this
report may have been blurred to protect their identities.
1.1 HISTORICAL INCIDENTS IN WHICH EGRESS WAS SIGNIFICANT
Although the World Trade Center building collapses are arguably the most significant building events
where building egress played a critical role, concern about the ability of occupants to escape from large
buildings is hardly new. Indeed, many earlier lessons were based on analyses of high-rise fires in New
York City. In 1911, the Triangle Shirtwaist Fire spread through the top three floors of a 10-story fire-
resistant building in New York (Fire Engineering 1977). The fire started in a corner of the eighth floor of
the building and quickly spread over the entire floor as well as the floors above by the windows, stairs,
and elevator shafts. There were 145 fatalities in the fire, all but one from the ninth floor of the building.
While many of the fatalities were located on the ninth floor, approximately 40 jumped from the building
to the street below to escape the flames, and another 10 perished when an exterior fire escape collapsed.
The fire was extinguished with hose lines from two standpipe risers in the stairwells and was under
control within 18 min. The upper three floors were a complete loss. Significant issues identified from the
fire investigation included the fact that there was limited access to the stairwells due to partially-blocked,
non-fireproof doors that opened inward, as well as exterior cast-iron fire escapes which loosened from the
This reference is to one of the companion documents from this Investigation. A list of these documents appears in the Preface
to this report.
NIST NCSTAR 1-7, WTC Investigation 1
Chapter 1
wall due to heat from the fire. Subsequent recommendations promoted fireproof egress stairways and
automatic sprinklers for buildings taller than 18 m (60 ft).
The Equitable Building fire in New York City in January 1912 is an early example of building collapse
following fire (NFPA 1912). The Equitable Building was a group of five linked buildings, the tallest
being 10 stories tall. Erected beginning in 1869, the buildings were constructed of so-called fireproof
construction with wood floors on brick or tile arches supported by wrought-iron and steel I-beams resting
on columns made mostly of cast iron. The fire originated in the basement of one of the buildings from a
discarded match and spread quickly to a tile-enclosed shaft containing two elevators and eleven small
dumbwaiters enclosed in wood. Within 45 minutes, the fire had spread throughout the upper floors of the
buildings and downward through numerous unprotected floor openings. Except for a few areas, the
building was completely gutted by the fire. Three separate sections of the building collapsed, with the
largest collapse involving all of the floors down to the basement on one side of the building. Since the fire
occurred before business hours, loss of life was limited to three employees on the upper floors and three
additional deaths attributed to collapse of cast-iron columns. Firefighter loss was limited to a single
fatality, as all personnel were ordered out of the building prior to the first collapse. Egress and firefighter
access was through a single continuous stairway from the basement level to the top floor, deemed
inadequate for escape in the subsequent investigative report. Recommendations included the need for
protection of floor openings, corridor partitions, and structural metal work, and inclusion of sprinklers in
all portions of office buildings where fire is most likely to occur. Two remote stairways enclosed in
fireproof shafts with fire doors at each floor were deemed necessary. Additional stairways were
recommended such that travel distance to a stairway was limited to 27 m (90 ft).
In 1945, a U.S. Army Air Force B-25 crashed into the Empire State Building in New York City resulting
in a significant fire on parts of the 78th and 79th floors from an estimated 3 m 3 (800 gal) of gasoline
sprayed from the plane crash (Hayne 1945). The crash and resulting fire caused 14 deaths and
approximately 25 injuries. The crash occurred on a Saturday morning when few building occupants were
present, and much of the office space surrounding the crash site was unoccupied. Several occupants of the
79th floor took refuge in a metal and glass partitioned office and were later rescued by the fire
department. According to the investigation report, the stairwells remained tenable throughout the incident
and provided fire department access and a safe means of egress for occupants of the upper floors not
involved with the initial gasoline fire. Fire department access was accomplished via elevator to the
65th floor and by stairwell the remaining 13 to 14 floors. The fire was extinguished approximately
35 minutes after the first fire department notification. Building design, timing of the fire on a Saturday
morning, and fire department response were credited with limiting the resulting damage and loss of life.
Important issues related to building egress identified in the investigation report include (1) limiting use of
elevators as a means of egress from upper floors, since the crash of one of the elevators to the sub-
basement might create apprehension of the dependability of the remaining elevators; (2) an understanding
that damage to stair shafts may be sufficient to prevent their use as a means of egress from the crash floor
and floors above; and (3) a realization that fire resistive building construction does not preclude damage
by fire involving building contents.
In August 1970, a fire at the 50-story One New York Plaza building extensively damaged the 33rd and
34th floors and spread significant smoke throughout the building (Powers 1970). The fire was first
detected in the concealed ceiling space of the 33rd floor and spread to exposed polystyrene insulation in
the south and west walls of the 33rd floor. The building was only partially occupied at the time of the
NIST NCSTAR 1-7, WTC Investigation
Background and Introduction
fire, with some of the floors above the fire unoccupied. Occupants evacuated either by elevators or down
the stairwells. Heavy smoke conditions were noted on many floors of the building. Two security guards
and two firefighters died from the fire, and 30 injuries resulted. Fire department access was accomplished
via elevators to the 30th floor and by stairs to the fire. The fire was controlled within 5 h. Reducing the
fire load of building contents, the need for automatic sprinkler systems, and the protection of steel
members by materials that cannot be readily removed or damaged were important issues identified from
the investigation. New York City Local Law No. 5, Fire Safety in High-Rise Buildings, resulted in large
part from a reaction to this and several other high rise-fires in New York at the time. Among other
provisions, Local Law 5 requires building compartmentation, with an exception for sprinklered spaces
(New York City 1973).
In February 1972 and February 1974, major high-rise fires occurred in Brazil, causing more than
200 casualties. In February 1972, a fire in the 31 -story Andraus Building in Sao Paulo resulted in 16
fatalities and more than 375 injuries (Willey 1972). The fire developed on four floors of a department
store and then spread up the exterior facade of the building, involving 28 floors of the building within
25 minutes. The fire gutted most areas of the building and damaged structural supports. The department
store occupied the lower seven floors above grade and was served by four open stairways and two
elevators. The remainder of the building was of office occupancy with a single 1 m (39 in.) wide
enclosed masonry spiral stairwell and five elevators. Door construction in the office stairwell was of
hollow-core, wood, or metal construction. Combustible interior finish and exterior facade were credited
for the rapid fire spread throughout all but the upper four floors of the building. It was reported that some
people used elevators to egress the building, while others used the single stairwell. Once a stairwell door
on the fifth floor failed, leaving the lower floors of the stairwell untenable, occupants fled toward the roof
of the building. Approximately 300 people reached the roof level heliport, while another 200 became
trapped in the stairwell. Rescue operations for those trapped in the stairwell included ladders from nearby
buildings on the fifteenth and sixteenth floor. The use of areas of refuge by nearly 500 occupants was
aided by stairway ventilation and wind velocity.
An unfortunately similar fire two years later, which started on the 12th floor of the 25 story Joelma
Building in Sao Paulo, and resulted in 179 deaths, 300 injuries, and total destruction of the building
contents (Sharry 1974). Inability of helicopters to rescue occupants trapped on the roof of the building,
inadequate means of egress from the building (a single 1.2 m (47 in.) unenclosed stairwell), lack of fire
protection, and presence of combustible contents within the building were noted as significant in the fire.
The majority of survivors of the fire made their escape through the use of the building's four elevators.
While this method was not recommended due to the possibility that occupants may become trapped, the
success of the evacuation was attributed to two factors: the use of elevator operators allowed the
elevators to be operated in an express mode (stopping only at desired floors), and the elevator power
supply was unaffected early in the fire.
In June 1989, a fire occurred in a 10-story office building in Atlanta, Georgia (Isner 1990). The Peachtree
25th building was an H-shaped building with two connected 10-story towers and a population of
approximately 1,500 people. Each tower measured approximately 76 m by 20 m, with the connection
measuring 21 m by 24 m. The fire began on the 6th floor of the south tower at approximately 10:30 a.m.
The ignition of the fire was attributed to an electrician working on an electrical switchbox. While the
worker was attempting to return power to a section of the floor by replacing a 200-ampere fuse, severe
arcing occurred. The arcing had sufficient energy to melt metal and ignite the interior- finish materials in
NISTNCSTAR 1-7, WTC Investigation
Chapter 1
the hallway. The electrician was severely injured and later died, although not as a direct result of the arc,
which was estimated as having lasted 60 seconds or more. The fire growth rate was extremely high, and
the fire spread was rapid. Multiple layers of wall covering promoted extraordinary fire spread rates,
which was not an unfamiliar fire hazard to fire investigators (Bouchard 1982; Demers 1980). The wall
coverings had completely burned out when the fire department arrived on the floor, only seven minutes
after notification. The intense black smoke quickly trapped about 40 occupants on the floor of origin.
Most occupants found a room and closed the door behind them, breaking out windows to vent incoming
smoke and waited to be rescued. At some point, one woman jumped from a 6th floor window and
sustained severe injuries. The fire department was not notified until an occupant of the building from a
remote floor activated a manual pull-station at approximately 10:30 a.m. Several occupants of the fire
floor were leaning out of a window in order to breathe when the fire department arrived on the scene at
approximately 10:34 a.m. Fourteen occupants were rescued via ladder truck, and 14 people were rescued
using the stairwells. In all, five people died because of this fire, the first multiple fatality high-rise office
building fire in the United States in 17 years (Isner 1990).
Several failure modes contributed significantly to the severity of the fire. There were no automatic
sprinklers, which allowed the fire to spread. The electrician did not follow proper procedure when
changing the fuse, resulting in the arc that ignited the wall linings and electrical equipment. The ignition
source was so severe that a fire in the electrical room was inevitable, however. Multiple-layer,
combustible interior- finishes also contributed to the rapid spread of the fire.
In many instances, these significant egress events resulting from fires in buildings have shaped building
codes requirements related to the egress system. Requirements for stairwell design, placement, and
capacity all evolved as a result of significant past fire incidents.
1.2 PREVIOUS FIRES AND EVACUATION INCIDENTS IN THE WORLD
TRADE CENTER
In February 1975, a fire in WTC 1 began on the 1 1th floor and ultimately spread from the 10th to the
19th floor extending through telephone closets on each floor (Powers 1975). Although not important
from an egress perspective, the fire provides an appropriate background for what occurred in the later
terrorist attacks in 1993 and 2001. The fire was initially reported by manual alarm at 11:35 p.m.
Automatic alarms from smoke detectors on the 11th floor through the 19th floor responded at about 1 min
intervals after the manual alarm. It was believed that the fire originated in an executive office on the
1 1th floor and spread to upper and lower floors through 0.30 m by 0.45 m (12 in. by 18 in.) openings in
the floors of utility closets on each floor. Four steel floor trusses were distorted slightly. Approximately
800 m 2 (9,000 ft 2 ) of the 11th floor was damaged, destroying about half of the contents and damaging the
remaining contents in this area. Virtually all combustibles, including fire retardant-treated wood paneling
on the telephone closet walls of the 10th and 12th floors, were destroyed. Limited quantities of
combustible furnishings on the 12th and 13th floors limited the spread of fires from the telephone closets
on these floors. Recommendations resulting from the fire included (1) provisions for automatic sprinklers
in areas where highly combustible material or large accumulations of combustibles are present, (2) the
installation of detectors in return air shafts on each floor to purge the return air and stop the supply of
fresh air to the fire area, and (3) fire stopping of all openings in floors or walls as well as in any wiring
installations. It was noted that sprayed fire-resistive material may not adhere properly to surfaces or may
be dislodged as other building services are installed.
NIST NCSTAR 1-7, WTC Investigation
Background and Introduction
On August 3, 1977, two Fuerzas Armadas de Liberacion Nacional (F.A.L.N.) terrorist bombs exploded in
midtown Manhattan, killing one person and injuring seven others. When a specific threat against the
World Trade Center was phoned into a local TV news station at 9:45 a.m. (Breasted 1977), both
WTC towers were evacuated, although not until after 12 noon. An employee of Windows on the World at
the time, described the situation:
"We were all scared. I started to shake. The ride down seemed to take
two hours. I'm part of a team that was trained for fire drills, but I have
no idea of what do if there was a bomb. This was more frightening than
a fire because we are all equipped for a fire." (Ivins 1977)
An estimated 35,000 people were evacuated from WTC 1 and WTC 2, both of which reopened the same
day, shortly after 3:00 p.m. (Ivins 1977). Overall, more than 100,000 people evacuated buildings in
Manhattan that day (Breasted 1977). Many people, however, were reluctant to leave after having been
docked wages after previous incidents for evacuating the building (Ivins 1977).
At 12:18 p.m. on February 26, 1993, a terrorist attack resulted in an explosion in a sublevel parking
garage in the World Trade Center complex, immediately killing six people (Isner and Klem 1993a; Isner
and Klem 1993b) and causing an estimated $300 million damage. The explosion of at least 450 kg
(1,000 lb) of explosive material caused extensive damage to several sublevels of the building and an
intense fire that spread varying amounts of smoke in four of the seven buildings in the complex. Most of
the complex's estimated 150,000 occupants evacuated the buildings as a result of the incident, including
approximately 50,000 from the affected towers. 2 According to the NFPA Investigation, 1,042 people
were injured in the incident, including 15 who received blast-related injuries. At the peak of the incident,
the fire reached 16 alarms and involved more than 700 firefighters (approximately 45 percent of the New
York City Fire Department's on-duty personnel) (Isner and Klem 1993a). As a comparison, on
September 11, 2001, 22 alarms were called prior to the collapse of WTC 2, in addition to a 10-60 alarm
(unique to special operations for large incidents) and a three alarm which staged additional units nearby.
This resulted in the involvement of more than 1,000 firefighters being at the World Trade Center.
The explosion significantly damaged floors, walls, and doorways in subgrade levels and forced large
amounts of smoke well away from the immediate area. In one report, visibility was reduced to 0.3 m
(1 ft) within about 1 min at the 44th floor of WTC 1, largely through the spread of smoke in elevator and
stairwell shafts (Isner and Klem 1993b). Before beginning evacuation, many occupants experienced
smoke on occupied floors and encountered even heavier smoke as they descended the buildings in the
stairwells. Since the explosion disabled the emergency communication systems in the buildings,
occupants responded to the event without the planned central guidance. Even without guidance, many
occupants began evacuation early in the event. Egress was further complicated by a total loss of electrical
power to emergency stairwell lighting within about 1 hour and 1 5 min. It was estimated that it took
occupants from IV2 hours to 3 hours to exit the building from the upper floors of the towers. Fortunately,
the scarcity of combustibles in the subgrade levels and dilution of the fire gases limited the toxic potency
of the resulting smoke. Although most of the injuries were smoke related, no fatalities due to smoke
inhalation were noted even with prolonged exposure to dense smoke.
Text of undated presentation by Ted Stam, General Property Manager, World Trade Center. WTCI-619-P.
NISTNCSTAR 1-7, WTC Investigation
Chapter 1
Fire crews were assigned responsibility for searching five floor subsectors. Since the elevators were not
operational, firefighters climbed the stairwells. It took more than two hours for crews to climb to the
100th floor. By 4:00 p.m., approximately 4 hours after the blast, all occupants had evacuated tenant
floors. Some elevator cars, however, had stopped in elevator shafts, portions of which had no doors
leading from that section of the shaft (such as express elevators). Locating and evacuating the trapped
elevator occupants was such a high priority that, rather than wait for the stairwells to empty to send the
elevator technicians up into the building by stairwells, the technicians were delivered to the roof by
helicopter. Even with that measure, it took approximately 5 hours to locate and free a group of
kindergarten students and several adults trapped in an elevator in WTC 2 (Isner and Klem 1993b). By
8:00 p.m., approximately 8 hours after the explosion, the last occupants trapped in elevators were
evacuated.
1.3 SCOPE OF OCCUPANT BEHAVIOR, EGRESS, AND EMERGENCY
COMMUNICATIONS PROJECT FOR THE NIST WORLD TRADE CENTER
INVESTIGATION
The purpose of this project was to determine the behavior and fate of occupants and responders - both
those who survived and those who did not - by collecting and analyzing information on occupant
behavior, human factors, egress, and emergency communications in WTC 1, WTC 2, and WTC 7, and
evaluating the performance of the evacuation system on September 11, 2001.
This project was divided into six tasks as follows:
Task 1 — Gathered baseline information on the evacuation of the WTC buildings on September 11, 2001,
through a comprehensive, systems-oriented, and interdisciplinary data collection effort focused on
occupant behavior, human factors, egress, and emergency communications (including instructions given,
interpretation of instructions, and response to instructions). This involved the collection of new data from
people affected by the WTC attacks (e.g. building occupants, building operators, and emergency
responders via direct accounts from survivors and families of victims), especially those who had to
evacuate the buildings. Experts in human behavior and statistical sampling were retained to assist in
developing a data acquisition strategy that considered various data collection methods, such as interviews
and questionnaires. Inputs and suggestions were obtained from individuals with an interest in the data
collection effort. Additionally, written accounts, transcripts of (emergency) communications, published
accounts, and other sources of egress related information were obtained, in coordination with other data
collection efforts for the overall investigation.
Task 2 — Collected archival records from prior WTC evacuation incidents (e.g., 1975 fire, 1977 blackout,
1980 bomb scare, 1990 power outage, and 1993 bombing) and practice evacuations, including oral history
data from floor wardens and fire safety directors. These records were compared and contrasted with the
September 11, 2001, evacuation. Changes made to the evacuation procedures following the earlier
incidents and in recent years were evaluated in the context of the experience on September 11, 2001.
Task 3 — Documented pre-event data for WTC Buildings 1, 2, and 7. This information included physical
aspects of building egress components, such as stairs (width, number, location, vertical continuity),
evacuation lighting, back-up power, elevators (number, operational before and after impact, role in
evacuation), and active fire protection systems (sprinklers, manual suppression, fire alarms, smoke
NIST NCSTAR 1-7, WTC Investigation
Background and Introduction
control). Building plans, emergency plans, type and frequency of evacuation drills, occupancy level and
distribution on the morning of September 1 1th, and communications also constituted pre-event data. This
information provided a baseline for evaluating the performance of the egress system.
Task 4 — Stored the information collected in task 1 in a database. Additionally, information from third-
party sources, such as published media accounts, were assembled and analyzed in the database.
Task 5 — Analyzed the data to study the movement of people during the evacuations, decision- making and
situational awareness, and issues concerning persons with disabilities. A timeline of the evacuation was
developed using the results of these analyses together with other data sources. This timeline was
compared with the timeline of the structural response, data on the development of the interior conditions
(fire and smoke), as well as information on the activation of the active fire protection systems. The
observed evacuation data was compared with results obtained using alternate egress models to better
understand occupant behavior and identify needed improvements to existing egress models. In addition,
the evacuation experience was compared with previous evacuation incidents in these buildings. The
results were reviewed in the context of occupant protection practices for tall buildings, including the
consideration of total evacuation and phased evacuation strategies.
Task 6 — Report preparation. The results of this project were synthesized into this report to describe the
occupant behavior, egress, and emergency communications in WTC 1 and WTC 2, and the performance
of the evacuation system.
1.4 REPORT ORGANIZATION
This report investigates the occupant behavior, egress, and emergency communications at the World
Trade Center on September 11, 2001.
This chapter explores historical fire incidents where egress played a significant role, as well as previous
significant fire or evacuation incidents at the World Trade Center complex. It also describes the scope of
the overall project.
Chapter 2 describes the design of the World Trade Center egress system, including the stairwells,
elevators and emergency communication systems. Emergency procedures, including the roles of building
managers is described. Finally, changes to the egress system as a result of the 1993 bombing are detailed.
Chapter 3 documents the overall technical approach of the project, including discussion of the collection
and analysis of first-person accounts (face-to-face, telephone, and focus group interviews), collection and
analysis of published media accounts, and collection and analysis of other relevant data, including audio,
video, photographic, and design records.
Chapters 4 through 9 chronologically detail the overall progression of the evacuation of WTC 1 and
WTC 2, including occupant activities, observations, and reactions.
Chapter 4 enumerates the occupants of WTC 1 and WTC 2, describes their basic characteristics as it
relates to evacuation, and discusses the emergency preparedness of the occupants prior to the attacks.
NISTNCSTAR 1-7, WTC Investigation
Chapter 1
Chapter 5 documents the occupants' awareness of and reaction to the impact of the first airplane with
WTC 1, as well as observations of local damage and phenomena.
Chapter 6 describes the period of time from immediately after WTC 1 was attacked until just prior to the
attack on WTC 2. The overall evacuation rate, actions of the building managers, and occupant activities
and behaviors are discussed.
Chapter 7, paralleling Chapter 5, documents the impact of the second airplane with WTC 2, including
awareness and reaction on the part of the occupants.
Chapter 8 tracks the progress of the evacuation and overall emergency response in both towers until
immediately prior to the collapse of WTC 2.
Chapter 9 examines the collapse of WTC 1 and WTC 2, including an analysis of where the occupants
likely were as each building collapsed.
Chapter 10 discusses the important egress issues raised by the events of September 11, 2001, at the World
Trade Center. Included are causal models, summary statistics on the overall building evacuation rates,
egress modeling, and in-depth analysis of specific issues that affected the evacuation, including the role of
alarms, announcements, mobility impairments, emergency responders, authorities, information flow,
activities, evacuation experience, and constraints/aids to evacuation.
Chapter 1 1 summarizes the key findings of this report and highlights the most important findings to
consider in response to the evacuation of WTC 1 and WTC 2.
NIST NCSTAR 1-7, WTC Investigation
Chapter 2
Design of the World Trade Center Egress System
The provision of access to and egress from buildings under emergency conditions relies on four primary
components: stairwells, elevators, communication systems, and emergency responders (broadly defined to
include the City of New York Fire Department (FDNY), New York City Police Department (NYPD),
Port Authority Police Department (PAPD), Port Authority personnel, building security, fire safety
directors, floor wardens, and other individuals with formal response responsibilities). These are
subsequently grouped into building systems and the human component.
2.1 OVERALL BUILDING DESCRIPTION
By 2001, the World Trade Center (WTC) complex had become an integral part of Manhattan. It was
composed of seven buildings (here referred to as WTC 1 through WTC 7) on a 1 6 acre site, located near
the southwest tip of the island, shown in Figure 2-1. Whether viewed from close up, from the Statue of
Liberty across the Upper Bay or from an airplane descending to LaGuardia Airport, the WTC towers were
a sight to behold. WTC 1 (often referred to as the North Tower) and WTC 2 (often referred to as the
South Tower), were each 110 stories high, dwarfing the other skyscrapers in lower Manhattan and
seemingly extending to all Manhattan the definition of "tall" set by midtown's Empire State Building.
Groundbreaking for the towers was in 1966, while construction began in 1968. WTC 1 was first occupied
in 1970; WTC 2 in 1972.
Additionally, there was a six-story subterranean structure, largely below the WTC Plaza with connections
to WTC 1, 2, 3, 4, 5, and 6, which included a shopping mall and the WTC PATH station. This was
surrounded by a 3 ft (0.9 m) thick concrete wall that extended from ground level down 70 ft (21 m) to
bedrock. Holding back the waters of the Hudson River, this wall had enabled rapid excavation for the
foundation and served to keep the groundwater from flooding the underground levels. Commuter trains
brought tens of thousands of workers and visitors to Manhattan from Brooklyn and New Jersey into the
WTC station. A series of escalators and elevators took the WTC employees directly to an underground
shopping mall and to the Concourse Level of both towers.
WTC 3 (Marriott Hotel) was 22 stories. WTC 4 (South Plaza Building) and WTC 5 (North Plaza
Building) were both 9-story office buildings. WTC 6 (U.S. Customs House) was an 8-story office
building. These six buildings were built around a 5 acre plaza, named for Austin J. Tobin, and the
centerpiece of which was a large globe art object. WTC 7, located north of the other six WTC buildings
and separated by Vesey Street, was a 47-story office building. WTC 7 was completed in 1987 and was
operated by Silverstein Properties, Inc., as an air rights building.
NISTNCSTAR 1-7, WTC Investigation
Chapter 2
Plaza Level
WTC 3 WTC 2
WTC4
Bridge to
130 Liberty Street
Liberty Street
Figure 2-1. WTC site plan.
2.1.1
Description of the Towers
WTC 1 and WTC 2 each consisted of 1 10 stories above the Concourse Level (or 109 stories above the
plaza / Mezzanine Level) structure. There were also six basement levels below the Concourse Level.
Although the towers were similar, they were not identical. The height of WTC 1 at the roof level was
1,368 ft (418 m) above the Concourse Level (6 ft taller than WTC 2), and WTC 1 additionally supported
a 360 ft (1 10 m) tall antenna on the roof for television and radio transmission. Each tower had a square
plan with the side dimension of 207 ft 2 in. (63.2 m). The corners of the tower were chamfered 6 ft 1 1 in.
(2.1 m). Each tower had a core service area of approximately 135 ft x 87 ft (41 m x 27 m), although the
core space changed on tenant spaces throughout the towers. A typical architectural floor plan in the tower
is shown in Figure 2-2. As can be seen in this figure, placing all service systems within the core provided
column-free floor space of roughly 31,000 sq ft (2,900 m 2 ) per floor outside the core. The long axis of the
10
NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
core in WTC 1 was oriented in the east-west direction while the long axis of the core in WTC 2 was
oriented in the north-south direct.
Figure 2-2. Typical WTC tower architectural floor plan.
The superb vistas from the top of such buildings virtually demanded public space from which to view
them, and the Port Authority responded. The 107th floor of WTC 1 housed a gourmet restaurant and bar
with views of the Hudson River and New Jersey to the west, the skyscrapers of midtown Manhattan to the
north, the East River and Queens to the east, the Statue of Liberty to the southwest, and the Atlantic
Ocean to the south. Similar views could be seen from observation decks on the 107th floor and the roof
of WTC 2.
NISTNCSTAR 1-7, WTC Investigation
11
Chapter 2
Table 2-1 shows the use of the floors, which was similar but not identical in the two towers:
Table 2-1. Use of floors in the WTC towers
Floor(s)
WTC1
WTC 2
Roof
Antenna space and window washing
equipment
Outdoor observation deck and window
washing equipment
110
Television studios
Mechanical equipment
108, 109
Mechanical equipment
Mechanical equipment
107
Windows on the World
Indoor observation deck
106
Catering
Tenant space
79 through 105
Tenant space
Tenant space
78
Skylobby, tenant space
Skylobby, tenant space
77
Tenant space
Tenant space
75,76
Mechanical equipment
Mechanical equipment
45 through 74
Tenant space
Tenant space
44
Skylobby, kitchen, tenant space
Skylobby, tenant space
43
Cafeteria
Tenant Cafeteria
41,42
Mechanical equipment
Mechanical equipment
9 through 40
Tenant space
Tenant space
7,8
Mechanical floors
Mechanical floors
Concourse through 6
6-story lobby
6-story lobby
The Port Authority had managed the operation of the two towers since their opening three decades earlier.
Silverstein Properties acquired a 99-year lease on the towers in July 2001.
At the beginning of the workday, many of the roughly 40,000 people who worked in the towers and
visited to tour or to conduct business emerged from PATH trains in the massive subterranean station.
They would take escalators and elevators to a large shopping concourse. Walking a few hundred feet led
occupants to the spacious, 6-story-high lobby on the Concourse Level where they would cross paths with
those who arrived on foot or by bus and cab. Figure 2-3 shows the layout of the shopping mall, located
underneath the WTC plaza. Figure 2-4 shows the lobby configuration for WTC 1. Figure 2-5 shows the
layout of the WTC 2 lobby. The WTC 1 and WTC 2 lobbies were at the same level as the underground
shopping mall, often collectively referred to as the Concourse Level. The WTC outdoor plaza and the
WTC 1 and WTC 2 Mezzanine were one story higher than the Concourse Level, often referred to as
either the Mezzanine or plaza level.
12
NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
Concourse (Mall) Area
Vesey Street
Li berty Street
Figure 2-3. Shopping mall layout underneath WTC plaza.
NISTNCSTAR 1-7, WTC Investigation
13
Chapter 2
Concourse
Figure 2-4. WTC 1 lobby (concourse) level.
14
NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
Concourse
Escalators
Down
Figure 2-5. WTC 2 lobby (concourse) level.
NISTNCSTAR 1-7, WTC Investigation
15
Chapter 2
Getting tens of thousands of people from the Concourse to their offices was no small task. This was
accomplished by a then-novel array of 106 express and local elevators located within the building core (as
shown later in Figure 2-14). Section 2.2.3 discusses the elevators system in WTC 1 and WTC 2.
Also within the core were three sets of stairs that extended the full height of the tower. Section 2.2.2
discusses the stairwells in each tower. However, upon entering a stairwell at an upper floor, one did not
find a continuously descending staircase leading to the lobby. Principally at the mechanical floors, there
were enclosed horizontal corridors that led around the massive elevator hardware. These corridors ranged
in length from about 1 ft to about 1 00 ft. After traversing each of these, the pedestrians would resume
their descent.
Upon exiting the elevators (or stairs, for those who chose the more strenuous route), one was faced with a
view typical of high-rise buildings. Surrounding the rectangular core corridor was a mixture of blank
walls, door entries to firms, and glass-front reception areas. Above was a standard drop ceiling.
Many of the floors had but a single tenant, and some of these tenants occupied multiple floors. By 2001,
most of these companies, which had moved in since the installation of automatic sprinklers, had taken
advantage of Yamasaki's design concept of a vast space that was virtually obstruction- free. The open
landscaping included as many as 200 or more individual workstations, often clustered in groups of six or
eight (Figure 2-6). Trading floors had arrays of long tables with multiple computer screens (Figure 2-7).
Some of these floors had a few executive offices in the corners and along the perimeter. Many also had
walled conference rooms. It was common for the multiple-floor tenants to have installed convenience
stairs internal to their space.
Other floors were subdivided to accommodate as many as 20 firms. Some of the smaller firms occupied
space in the core area, reclaimed as elevator shaft space from local elevators was phased out throughout
the towers.
With so many workers and visitors in the buildings, there needed to be food available. The Port
Authority maintained a cafeteria on the 43rd floor of WTC 1. A number of the companies maintained
kitchen areas where catered food was brought in daily, making it unnecessary for their staff even to leave
their floor for lunch. The underground Concourse Level mall also provided may options for eating. In
addition, there were hundreds of restrooms, in both the tenant and the core spaces.
1 6 NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
Figure 2-6. Typical WTC tenant
spaces.
Source: Photos courtesy of The Port Authority of
New York and New Jersey.
Figure 2-7. A WTC 4 trading floor.
Source: Photo courtesy of The Port Authority of
New York and New Jersey.
NISTNCSTAR 1-7, WTC Investigation
17
Chapter 2
2.2 BUILDING SYSTEMS
2.2.1 Egress Calculations
Determining the Number and Width of Stairwells
In 1965, architects and engineers designing the World Trade Center towers were faced with an impending
change to the NYC Building Code. The draft building code language had a significant impact on the
design of emergency egress systems. In 1965, the Port Authority directed its designers to adopt the draft
version of the new code for their final designs. Some of the advantages of the new draft code were noted
to be the following (Levy 1965):
• Fire towers 3 could be eliminated;
• Provisions for exit stairs were more "lenient;" and
• Criteria for partition weights were more "realistic."
It was not certain whether all the changes being proposed to the 1938 code would be incorporated into the
final version of the new code. Thus, in 1966, the Chief Engineer of the Port Authority suggested that the
"architect/engineers prepare a listing of the elements of the design which do not conform to old code
requirements, but are acceptable under the new. With this list in hand, we could initiate discussions, at
top level in the Building Department, to see if we can secure agreement to go along with our design"
(Kyle 1966).
A one-page document, dated "2/15/67", with the initials "CKP" listed the following items: 4
• Fire tower corridors [sic] eliminated.
• Number of stairs reduced from 6 to 3. (Old plans had 5 stairs at 3'-8" and 1 stair at 4'-8" for a
total population of 390. 5 New plans have 2 stairs at 3'-8" and 1 stair at 4'-8" allowing a
population of 390.)
• The size of doors leading to the stairs are [sic] changed from 3'-8" to 3'-0".
• All stairs exit through a lobby. Old plans had fire tower stair exiting through a fire enclosed
corridor.
• Shaft walls are changed from a 3 h rating to a 2 h rating.
• Corridors are limited to a 100 ft dead end and with a 2 h rating.
A fire tower is an exterior stairwell of incombustible construction terminating at grade level designed to ensure that smoke
conditions from an interior fire do not contaminate the fire tower. The fire tower was provided for firefighter ingress and did
not count as a required stairwell for occupant egress. (NYC Building Code 1938)
See appendix of NCSTAR 1-1 for a reproduction of this memo.
5 The 1938 NYC Building Code allowed 30 person per unit of exit width, while the 1968 NYC Building Code allowed
60 persons per unit of exit width, effectively halving the egress capacity of new construction. Population calculations are
per floor.
18 NIST NCSTAR 1 - 7, WTC Investigation
Design of the World Trade Center Egress System
• Additional (word(s) missing) changed from 20 pounds per ft 2 to 6 pounds per ft 2 (based on
partition weight of 50 pounds to 100 pounds per linear foot).
Apparently, this list represented elements of the WTC design that would not have satisfied the 1938 code,
but did satisfy the then-current draft version of the new code. Ultimately, WTC 1 and WTC 2 were
designed with three stairwells, two 3 '-8" (44 in.) wide and one 4'-8" (56 in.) wide, as discussed below.
A unit of exit width in the 1968 NYC Building Code was (and continues to be) 22 in. (0.56 m). The NYC
Building Code table specifying exit and access requirements (Table 6-1) required that for a business
occupancy, the stairs would accommodate 60 persons per unit of exit width. As the WTC 1 and WTC 2
tenant floor design occupancy load was 365 persons per floor 6 (Solomon 1968), this required 6.5 units of
exit width. Twelve in. (0.3 m) was the minimum half-width acceptable in the code, therefore, three
stairwells (two with two units of exit width (44 in. [1.1 m]) and one with 2.5 units (56 in. [1.4 m]))
satisfied the minimum requirements of the 1968 NYC Building Code. Table 2-2 shows the location of
the stairwells, core perimeter, and transfer hallways for occupied floors in WTC 1 and WTC 2.
Egress Provisions from Windows on the World
The 106th and 107th floors of WTC 1 (North Tower) contained the Windows on the World complex,
consisting of the Windows on the World restaurant, the Greatest Bar on Earth, numerous banquet and
function rooms, kitchens and support areas, and management offices for the dining complex. While the
configuration of the space may have changed over the life of the building, these functions were all present
from the time Windows on the World first opened in April 1976. 7
Restaurants, bars, and function rooms are classified in building codes as assembly use, which carries a
significant increase in occupant load and consequent provisions for egress. The design occupant load for
such assembly space is 1 5 ft 2 per occupant as opposed to the 1 00 ft 2 per occupant for the office use space
in most of the rest of the buildings. Thus, while the design number of occupants on an office floor was
365 to 390 (depending on the calculation method), the design number of occupants for these floors was
over 1,000 each (the exact number depends on the area of kitchens, dishwashing, and office space on the
floor, all of which is at 100 ft 2 per occupant).
Locating assembly space high in a building poses particular challenges to egress design because the
capacity of an egress component is not permitted to be decreased in the direction of travel. Thus, where
more or wider stairs are provided to meet capacity requirements these must be continued all the way
down through the building which affects space utilization for the entire structure.
A January 25, 1968 memo from J. Solomon (Emory Roth and Sons) to M. Levy (PANYA) subsequent to a NYC Building
Department plan review, documents that the "largest floor area is about 36,500 ft on the 106th floor. At one person per
100 ft there will be 365 persons per floor, well within the permissible maximum" of 390 persons based upon stairwell
capacity. WTCI-477-P Note that this calculation did not account for the use of the 106 th floor as an assembly space.
7 PANYNJ response to formal NIST question, March 25, 2005.
NIST NCSTAR 1-7, WTC Investigation 1 9
Chapter 2
Table 2-2. Plan view of stairwells in WTC 1 and WTC 2.
.oi^ 1 ""
1,111111"'
,#«>
• ^\
WTC 2
WTC1
I--
#l ^<^ N
-.'«"
©
WTC 2
V"^
WTC1
Lobby (Concourse) Level Layout
Only Stairwell B Serviced the Concourse Level
MST
X
>v
\
&
©
WTC 2
\
f0*'
©
wrci
Mezzanine (Plaza) Level Layout
Only Stairwell B Serviced the Concourse Level
WTC 2
WTC1
Floors 9-19
Stairwells A was East in WTC 1, North in WTC 2.
WTC 2
WTC1
Floors 20 - 26
Core space previously used by the elevators in the northeast
quadrant became leasable tenant space.
tf##
„\»»'
,«i»'
WTC 2
Ull"'
,1111"
mill'"'
,,,,,,1111 %
"•«,
<± &
WTC1
Floors 27 - 34
Floors 35 - 41
20
NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
1-."—
x
\
t
^
©
WTC2
\
10
©
WTCl
)|||fl "Hi,
C^
•O
\
^
^
©
V
^
©
WTC2
n
W
WTC1
Floor 42
Stairs A and C transferred outside the core.
^
***
^.--v
- ^
\
S
t u
^
©
\
tfH0^
©
WTC2
II
w^
WTCl
Floors 43, 45 - 47
There was an escalator connecting floors 43-44 (skylobby).
»!!
,,11111'"""
WTC2
WTCl
Floor 44 (Skylobby)
Floor 48
Stairs A and C transferred back inside the core.
#""" "X,
WTC2
,^«""" /'X
WTCl
WTC2
WTCl
Floors 49 - 54
Floors 55 - 56
NISTNCSTAR 1-7, WTC Investigation
21
Chapter 2
^
WTC2
WTC1
Floors 57 - 75
There was a slight change in Stairs A, C between floors 66-68.
Floor 76
Stairs A, B, and C transferred, with Stairwells A and C moved
outside the core.
S \0^
©
WTC1
■*■
<<■
>
WTC2
*
©
WTC1
Floors 77, 79, 80, 81
There was an escalator connecting floors 77-78 (skylobby).
Floor 78
WTC2
WTC2
WTC1
Floor 82
Stairs A and C transferred back inside the core.
Floors 83 - 95
22
NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
,„,,..-"
r o% ^
©
WTC2
\X
WTCl
Floors 96 - 102
ram
&
\
V
<s>
<>
X
\
V
©
^kama
,;:::;:#
©
WTC2
w
F^
WTCl
Floors 103 - 104
x
© N
WTC2
V^
©
WTCl
Floors 105 - 106
Floor 107
The document record contains a letter dated January 27, 1995, from Eugene Fasullo (PANYNJ) to
Richard Visconti (Deputy Commissioner, NYC Department of Buildings [DOB]) confirming the results
of a meeting on December 6, 1994, at which they reached agreement on a plan to address egress
requirements from the 106th and 107th floors (Fasullo 1995). The details of the agreed solution are
summarized below. The Deputy Commissioner, DOB, signed the letter to show concurrence with the
agreed solution. 8
It remains unclear what conditions existed from the date Windows on the World first opened to the time
the agreed solution was implemented in 1995. The dates suggest that the issue was identified as a result
of the Memoranda of Understanding (MOU) between PANYNJ and the NYC DOB and FDNY executed
in 1993, in response to the bombing. A Windows on the World refurbishment after the 1993 bombing
included these egress system changes. 9
The basis for the agreed solution was to divide each floor into three areas of refuge (consistent with
Section 27-372 [NYC Building Code]) to provide additional capacity to the existing stairs in accordance
Fasullo, E., PANYNJ, to R. Visconti, NYC Department of Buildings, "Variance Granted by Memorandum of Understanding
with Buildings Department, Windows on the World," January 27, 1995.
PANYNJ response to formal NIST question, March 25, 2005.
NISTNCSTAR 1-7, WTC Investigation
23
Chapter 2
with Section 27-367 (NYC Building Code). These identical provisions existed in the version of the
1968 NYC Building Code in effect when the buildings were built as sections C26-604.5 and C26-603.3,
respectively (the NYC Building Code was renumbered) (NYCBC 1968).
These code provisions allow for a doubling of allowed stair capacity when one area of refuge is provided
on a floor and tripling the stair capacity for two or more areas of refuge on a floor. These areas of refuge
must be separated by 2 h construction, be large enough for the expected occupant load at 3 ft 2 per
occupant, each contain at least one stair, and have access to at least one elevator (above the 1 1th floor).
Since three, distinct areas of refuge were provided on each floor, tripling of the capacity of each of the
three stairs resulted in a maximum permitted occupant load of 1,170 people per floor (6.5 units of egress x
60 persons per unit x 3).
Attached to (and referenced in) the letter were two plans entitled "106th Floor Egress Plan" and "107th
Floor Egress Plan" (shown in Figure 2-8 and Figure 2-9, respectively) that detailed the arrangement,.
The 2 h separation walls snaked across the floors and were not aligned on the two floors. Some areas that
needed to remain open to free passage were protected with Won doors (accordion doors that are fire rated
and are closed automatically on activation of the fire alarm system). Details of the egress system design
calculations and corresponding NYC Building Code requirements were included on the plans to
demonstrate they met code requirements.
By comparison, current model building codes, including the ICC 2003 International Building Code (IBC)
and National Fire Protection Association (NFPA) 5000, both permit a doubling (but not tripling except in
IBC Type 1-2 and 1-3 institutional uses) of the stair capacity for the provision of a horizontal exit on a
floor. The horizontal exit must consist of a 2 h fire rated separation, contain at least one stair on each
side, and have sufficient space for the expected occupant load at 3 ft 2 per person. A horizontal exit must
be continuous down through the building to grade 10 (NFPA 1 1.2.4.3.1 and IBC 1021.2), unless the floor
assemblies are at least 2 h with no unprotected openings.
The solution to the egress problem was to provide a protected space in which occupants could wait to
enter stairs that did not have adequate capacity for the numbers of people. Since the attacks took place in
the morning (a non-peak time), NIST estimated that there were 188 occupants trapped in the Windows on
the World floors. If the attacks had occurred when the facility was loaded near its capacity, as many as
2,000 occupants could have lost their lives on those two floors alone, since there were no survivors above
the impact floors of WTC 1 .
In other words, the stairway may not contain unprotected openings (such as opening out to a floor) until the occupant exits the
building.
24 NIST NCSTAR 1 - 7, WTC Investigation
Design of the World Trade Center Egress System
III
Be
2:
:.> .-rrnr.'-T J :-.- y, f. ,,-- ■
i-i t' 'it r* »i "• i
l'jJ ,e JL&.-J j, |r ..^n- r
' '•* r ' r is & ~ ' ■ ' i i * * : * ■
■ fij 4 ^ >! *tj ■!■ !l
. ^_-. ,■-'.'. '^--^-.., .-.----■- _-w_..
„ . . ■ v i j j : j : . .
I ■
llffiM
Hi , III V^tl
nil tt Jii -4
-
4 S
™ tn
ii Ml
UJ o o
a: = e
= m-
in irt ll
<N "— £>
II » U
is S m
woo
ljj n q
Q O S?
@:
CV ^ .&
It ii "o
^ li (0
's 8 m
to o o
ft °-
Ul is
""- CM
" II
3 2^
go e
I J £
U) o o
tO
S.S,
■SS
IMS
_ OL "- m
h^ C SC TT
[^ i^ rt> a
@Sg s
■g in
o
c
CO
"o
c
o
Q_
w
£
s_
b
+*
o
Q.
c
o
-£=
o
Li.
g
5
o
■o
(D
O
O
cu
o
o
L.
Q.
c
"*~~
cr
I* 5= b
(O
o
r-
rH
U)
^_
^
o
>s
o
-£-!
4—
!_
<1>
c
F
o
CD
+->
r
CC
cc
-J
o
LU
CS
o
<1)
en
0)
en
(1)
u
■-J
TO
CD
^:
T3
"O
i-
!_
O
o
(1)
>
-C
^
+J
CD
r
^
o
O
en
^
§
o
r"
o
T3
Z3
C
2 8
cu at
oo
i
CM
CD
;_
3
0)
o
3
O
CO
NISTNCSTAR 1-7, WTC Investigation
25
Chapter 2
f 4~
CD
LL <
.6 I
£ *
IF
r_j
O-
s
CO
D-
E>5
O,
m S
*- Si
cmj
c
£
u
ai
to
A3
cz
&
^
£
>l =)
II 3
11 II
BE!
D_
^ UJ <£
- } m £
Oi
M
■:
1
en
g§ 1
&g
o
LO
C
n
.5
to 3
0)g
ID (jgQ.
tO @CL
©*"
a.
en
n
L'-
©=■8
@ S R
3 <B>
35
CM
£
O
g*
= UJ —
LU 3
CD
---
LU in ii
LU lC. II
in LU
II
m
■
m
CM *~~ ^*
CM ■"- £?
cm CM
ft 1
11
ii II 'O
■5^3.
II H tj
■5 ° ^
ii 11
>-
s
*
5 o t
9 o to
p.-.
^3
c
<
8.
to
1
CO Q O
6
d
£j0 a
l*i
d
(J
TJ
Zi
Q
3.
T3 "■
™§
— c
a) £
5 e
e s
[1
: 1
5£
I ij
<
fl
i
1
3
3
^ JS K
M
ff
a? g>
£
dE
e
€
n
ED
in
S s
s
s
SR
ia
—i
01
J u.
■CM
■ri . L ;
3
S
a s
ii
1
«!
4)
SM
^
ft
n%z
Ql
S.HJ
C3
UJ^
S.S I
O
r/l
" 3)
U5 " 5)
a
en
LJ
fl
o
F
-
Q_ :ii
US
y
■i^
1TJ
a>
<
Q-
< O A3
m
I
^ ni
s
cm £
,-'
U- P
rt u_ P
E
^
Q
O
o
^
K
2
2^
>n
jB
JU
3S
E
F
E
V
q
9
*o
1
I s
i
s
e
"5
■5
"5
S
s
S
z
o_
D_
fl_
CD D_
*■ O
c o
to -
^^
° ?;
r»
o
H
rH
(/)
^_
^
o
>>
o
-£J
4—
!_
(1)
c
1=
o
QJ
O
C
CC
CC
-J
o
LJJ
CC
o
(1)
(fl
(fl
a)
,_ )
1—
3
0)
TO
^
T3
"□
i-
O
^
5
o
>-
u
£
■5
■l-l
0)
r
^
o
o
(A
^
§
o
r
o
T3
-I
C
<
g
o
CO
LL
O)
CU
O)
J_
rH
H
O
O)
C
1
o
CM
in
a)
b
"5
fl)
TO
LL
r
LL
g
T3
ri)
()
1
T5
fl
!—
n
CI)
tr
cu
o
!~
3
O
w
26
N/ST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
A similar condition existed on the 107' floor of WTC 2, commonly referred to as the Observation Deck.
A tenant alteration application submitted by Ogden Entertainment (the tenant) to PANYNJ in late 1 995
and early 1 996 utilized the areas of refuge provisions referred to previously with respect to the Windows
on the World space. ' Taking advantage of a NYC Building Code provision which permits a lower basis
for occupant load, the PANYNJ permitted a maximum occupant load of 1,170 persons on the floor
(Indoor Observation Deck and Outdoor Observation Deck, combined), which was enforced by the lessee
with periodic oversight by PANYNJ. 12
2.2.2
Stairwells
WTC 1 and WTC 2 each had three
primary stairwells designed for
emergency egress, designated as A, B,
and C. There were additional
stairwells located in the basement
levels (B 1 - B5), convenience stairs
for tenants leasing multiple floors,
and mechanical room stairs. These
secondary stairs are not considered
part of the emergency egress system
and are not described here. Stairwells
A and C were 1.1m (44 in.) wide and
extended from floor 2 (plaza or
Mezzanine Level) to floor 110 (lower
mechanical space). The stairwell
landings by the exit door were 92 in
(2.3 m) wide by 78 in (2.0 m) deep.
Figure 2-10 shows a 44 in. (1.1 m)
stairwell in WTC 1 taken on September
11, 2001, by John Labriola during his evacuation. Note the photo luminescent paint on the stair edge and
landing. Stairwell B was 56 in. (1.4 m) wide and ran from the subgrade 6 levels below ground to floor
107 including the Concourse (main lobby); there was no exit from Stairwell B onto the 2nd floor (plaza /
Mezzanine Level). The stairwell landings by the exit door for Stairwell B were 1 16 in (2.9 m) wide by 78
in (2.0 m) deep.
The 1968 NYC Building Code has requirements for the number and capacity of stairs and for the assumed
occupant load that are similar to requirements in the other contemporaneous codes (see NIST
NCSTAR 1-1, Appendix A). Codes of that time required that multiple stairs be located "as remote from
each other as practicable." NYC permitted scissor stairs, 1 and the code required the exit doors to be at
least 4.6 m (15 ft) apart. Local Law 16 (1984) first imposed a remoteness requirement of 30 ft or one-
Figure 2-10. 44 in. stairwell in WTC 1 taken on
September 11, 2001.
Ogden Entertainment. 1996. Port Authority work number W96-2103-01. WTCI-180-P.
12 PANYNJ response to formal NIST question. "Re: Question for PA." March 25, 2005. S. Bohl to S. Sunder.
Scissor stairs are two separate stairwells with two separate stairwell access doors, which share a common shaft space, often
winding around each other. This results in an efficient use of space, but places the stairwells in direct contact (in other words,
there is not a barrier separating the stairwells), thus allowing smoke or other threats to affect two stairwells simultaneously.
NIST NCSTAR 1-7, WTC Investigation
27
Chapter 2
third the maximum travel distance of the floor (whichever is greater). This requirement was not
retroactive, so it did not apply to WTC 1 and WTC 2. However, this requirement did apply to WTC 7.
The 1968 NYC Building Code also states that, " ...vertical exits should extend in a continuous enclosure
to discharge directly to an exterior space or at a yard, court, exit passageway or street floor lobby ..."
(C26-602.4). The 1965 BOCA Basic Building Code and 1966 NFPA 101 contained similar language, but
not the 1964 New York State Building Construction Code or the 1966 Municipal Code of Chicago.
Current model code language (2003 IBC, section 1003.6) defines continuous as: not "... interrupted by
any building element other than a means of egress component."
The exit discharge language was the subject of discussion in that the stairs in WTC 1 and WTC 2
discharged onto the Mezzanine Level, which was not at street level but rather at the Plaza level. The Port
Authority took the position that the concourse was like an underground street, and the arrangement met
the intent of the Code, as demonstrated by a February 18, 1975 letter from Joseph Solomon (Emory Roth
and Sons) to Malcolm Levy (PANYNJ), which covered six points. "We [Emory Roth and Sons] were
instructed by the Port Authority to deviate from the code [1968 NYC Building Code]." The fourth point
listed the "treatment of concourse level as 'Underground Street' noted by letter to the Port Authority on
April 6, 1971, January 11, 1972, and May 7, 1973" (Solomon 1975).
Transfer Hallways
The WTC 1 and WTC 2 stairwells were occasionally routed horizontally around equipment on
mechanical floors, through what were called transfer hallways, as shown in Figure 2-11. Table 2-2
shows the overall layout of the stairwells in WTC 1 and WTC 2, including the basic core perimeter. 14
Stairwell B required a horizontal transfer at floor 76. For all other floors, stairwell B maintained vertical
alignment through the building. Stairwells A and C required horizontal transfers (some longer than
others) at floors 42, 48, 66, 68, 76, and 82. Horizontal transfer distances ranged from several feet
(floors 66 and 68) to over 100ft (33 m), including smoke doors (which were closed but not locked) and
multiple right angles turns in the transfer on floors 42, 48, 76, and 82 for Stairwells A and C. Note that
the mechanical floors were located on floors 41-42, 75-76, and 108-109. One problem with the horizontal
transfers was that they extended the total evacuation time, when compared to a similar design without
horizontal transfers. The World Trade Center Review Committee, formed by the New York City
Building and Fire Commissioners in response to the 1993 WTC Bombing, found that "the occupants of
the towers encountered changes in the path of egress that were unfamiliar, [contributing] to the general
confusion during the evacuation process (New York City 1995)." Figure 2-12 shows a photograph of a
horizontal transfer hallway in WTC 1 or WTC 2 taken after the 1993 bombing, including
photoluminescent markings.
Core is defined in this report as the boundary of non-leasable common space, including egress hallways, stairwells, elevator
shafts and lobbies, HVAC, plumbing and other mechanical spaces. This definition of core may differ from a structural
definition of core, defined by the location of core columns, which did not change location in WTC 1 or WTC 2 on different
floors.
28 NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
WTC 2
Figure 2-11. Stairwells in the WTC towers.
WTC1
NISTNCSTAR 1-7, WTC Investigation
29
Chapter 2
Each stairwell had signage on both
sides of the stairwell access doors
indicating the letter designation of the
particular stairwell. A sign on the
inside of the stairwell indicated the
floor number, the stairwell
designation, and whether the floor was
a "re-entry" or "non-re-entry" floor.
Figure 2-13 shows a photograph of
this signage taken after the 1 993
bombing. A non-re-entry floor was a
landing in the stairwell where the door
to the floor was locked from the
stairwell side. If the particular floor
was not a re-entry floor, the sign
indicated the location of the nearest re-
Figure 2-12. Horizontal transfer floors in the
WTC towers.
entry location, every fourth floor (in the case of
Figure 2-13, floors 74 and 78). The stairwell doors
were required to be always open every fourth floor by
the NYC Building Code. Door locks leading to
mechanical spaces and the roof were controlled
electronically at the Security Command Center (SCC)
on floor 22. The NYC Building Code also required
that, in the event of a power outage, the re-entry
locking mechanism would default to the open position.
Compartmentation
The design of WTC 1 and WTC 2 featured large, open
office spaces devoid of columns due to the innovative
structural design. Tenants could (and often did) utilize
open plan office layouts that permitted impressive
views of the Manhattan skyline out the perimeter
windows.
The NYC Building Code and PANYNJ practice
required partitions to separate tenant spaces from one
another and from common spaces such as the corridors
that served the elevators, stairs, and other common
spaces in the building core. Fire rated partitions are
intended to limit fire spread on a floor and to prevent
spread of fire in one tenant space to that of another.
Partitions separating tenant space from exit access
corridors were permitted to be 1 h, although PANYNJ
specified them to be 2 h (Kyle 1966). This allowed dead ends to extend to 100 ft (rather than 50 ft with
1 h partitions), which permitted more flexibility in tenant layouts. Partitions separating tenant spaces (so
Figure 2-13. Stairwell door signage as
seen from inside stairwell A.
30
NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
called demising walls) were required to be 1 h. Enclosures for vertical shafts, including stairways and
transfer corridors, elevator hoistways, and mechanical or utility shafts were required to be of 2 h fire rated
construction. Protection of vertical shafts was intended to limit the spread of fire and smoke from floor to
floor.
Another influence on compartmentation of the buildings was the adoption of Local Law 5
(New York 1973) (LL 5) amending the NYC Building Code. While it did not (legally) apply to the
WTC buildings, PANYNJ policy was to follow the requirements voluntarily. LL 5 required
compartmentation of unsprinklered spaces in existing office buildings over 100 ft in height "having air-
conditioning and/or mechanical ventilation systems that serve more than the floor on which the equipment
is located" to be subdivided by 1 h fire separations into spaces or compartments not to exceed 7,500 ft 2 .
Floor areas could be increased up to 15,000 ft 2 if protected by 2 h fire resistive construction and smoke
detectors. Regardless of the floor area, compartmentation is not required when complete sprinkler
protection is provided (LL 5, Section 6).
Shortly after the adoption of LL 5, PANYNJ began to add the required compartmentation as a part of new
tenant layouts as evidenced by several subsequent tenant alteration contracts from this time. Following
the 1975 fire a fire safety consultant report recommended to PANYNJ that the buildings be retrofitted
with sprinklers to address possible smoke problems, which would also obviate the need for
compartmentation and permit the unobstructed views for which the buildings were known. The decision
left the interior WTC floor arrangements with only partitions separating tenant spaces from one other and
from exit access corridors or common spaces in the core, and with shaft enclosures. 5
Construction of Partitions and Shaft Enclosures
Vertical shafts surrounding stairs, mechanical shafts (carrying supply and return air), elevator hoistways,
and utility shafts were all contained within the building core of the WTC towers, and were enclosed by
gypsum planking similar to fire separations commonly used today in single-family attached housing.
These gypsum planks were 2 in. thick and 2 ft wide, reportedly with metal tongue and groove channels
attached to the long sides. These were likely two 1 in. panels held together by the metal channels. Their
length in WTC 1 and WTC 2 is unknown, but similar panels today are available in 8 to 14 foot lengths.
The planks were placed into metal H-channels at the top and bottom and secured by drywall screws.
The 1978 edition of the Gypsum Association (GA) Fire Resistance Design Manual lists several similar
shaft wall constructions utilizing 2 in. gypsum planks consisting of two 1 in. gypsum core board panels
with "metal channels on long edges." The GA Manual lists shaft walls of a single 2 in. metal edged plank
(WP7015) having a 1 h fire rating, a single 2 in. metal edged plank with one layer of Type X gypsum
board on the unexposed side (WP71 12) having a 2 h fire rating, and a single 2 in. metal edged plank with
two layers of Type X gypsum board on the unexposed side (WP 7575) having a 3 h fire rating.
Partitions separating tenant spaces on the same floor were constructed of two layers of 5/8 in. Type X
gypsum board on steel studs and ran slab to slab. This construction is commonly recognized as a 2 h fire
separation. Above the ceiling, penetrations for ducts or to allow for return air flow were fitted with rated
fire dampers to preserve the fire rating.
15 PONYA 1976 - Complete report reproduced in NIST NCSTAR 1-1H.
NIST NCSTAR 1-7, WTC Investigation 3 1
Chapter 2
Interior partitions not separating spaces occupied by different tenants were constructed of single or double
layers of 5/8 in. Type X gypsum board on steel studs, and ran from the slab to the suspended ceiling but
not above. Double layers of gypsum board were used when the tenant desired additional sound
attenuation. These partitions were not required to be fire rated and fire rated doors were not used.
However, a single layer of 5/8 in. Type X gypsum board on steel studs (16 in. on center) is generally
considered to have a 1 h fire rating and two layers of 5/8 in. Type X gypsum on steel studs (16 in. on
center) is considered to have a 2 h fire rating. For a ceiling-high partition to be considered as having a
fire rating, the ceiling itself would have to be rated as well. The ceiling system used throughout these
buildings was not fire rated.
2.2.3 Elevators
Getting thousands of people from the ground level to the offices, observation levels, and restaurants, some
as high as a quarter- mile, was no small task. Thus, elevators were the primary mode of movement
between floors of the World Trade Center. The World Trade Center complex contained more than
240 elevators, with 99 elevators serving the above-ground levels in each of the two main towers and an
additional 7 elevators serving primarily the sub-grade basement levels. In the towers, the elevators were
arranged to serve the buildings in three sections divided by skylobbies, which served to distribute
passengers among express and local elevators. Figure 2-14 shows an elevator riser diagram for WTC 1
and WTC 2 for passenger elevators.
• People traveling to floors 9 through 40 entered a bank of 24 local elevators at the Concourse
Level. These were divided into four groups, with each stopping at a different set of eight or nine
floors (9 through 16, 17 through 24, 25 through 31, and 32 through 40).
•
•
Those going to floors 44 through 74 took one of eight express elevators to the 44th floor skylobby
before transferring to one of 24 local elevators. These 24 were stacked on top of the lower bank
of 24, providing additional transport without increasing the occupied floor space.
Those going to floors 78 through 107 took one of 10 express elevators from the Concourse Level
to the 78th floor before transferring to one of 24 local elevators. These were also stacked on the
lower banks of 24.
Dedicated express elevators served the restaurant, bars, and meeting rooms on floors 106 and 107
of WTC 1, as well as the observation deck in WTC 2.
An occupant traveling to the 91st floor, for example, would have taken an express elevator from the lobby
to the 78th floor and then would have had to transfer to another elevator to arrive at the 91st floor. The
elevator trip would have taken several minutes travel time, depending upon the wait at the elevators.
While providing an acceptable rate of people movement, this three-tier system also used less of the
building footprint than the usual systems in which all elevators run from the entrance to the top of the
building. Further, leasable floor space was reclaimed near the top of a given zone. At the top of each
elevator bank, the machinery to lift the cabs occupied the next higher floor. From the next higher floor up
to the bottom of the next elevator bank, there was no need for an elevator shaft. The concrete floor was
extended into this space, providing additional rentable floor area for offices, conference rooms, storage,
etc. Figure 2-14, for example, shows that the space taken by Elevator Bank A (Elevators 24 - 29) in
32 NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
order to serve floor 9 to floor 16, was reclaimed for tenant use on floors 19 to 42. This resulted in a
reclamation of approximately 750 ft 2 per floor. A calculation for reclaiming unused floor space above
elevator banks A, B, and C for all three zones, reveals that roughly 100,000 ft 2 of potentially leasable
office space could be recovered. Assuming $55 per ft 2 per year as a rental rate for a downtown
Manhattan office building over 600,000 ft 2 (BOMA 2001), the reclamation could theoretically yield
nearly $6 million per year of rental income. At the time WTC was built, the concept of skylobbies,
served by express elevators and serving only one zone of the building, was innovative. Other tall
buildings now use this concept.
Mechanical Equipment -p.
Room
Skylobby -•_
Mechanical Equipment -•
Room =
Skylobby -J|
Mechanical Equipment
Room
Express Elevators - =
Mechanical Equipment
Room
Plaza Level
: =
Local Elevators
Express Elevators
Local Elevators
Figure 2-14. Elevator riser diagram for WTC 1 and WTC 2.
NISTNCSTAR 1-7, WTC Investigation
33
Chapter 2
In addition to the passenger elevators, there were seven freight elevators in each tower; most served a
particular zone, while Car 50 served every floor.
Car #5: B 1-5, 7, 9-40, 44
Car #6: Bl-5, 44, 75, 77-107 (Dual-use express, see below)
Car#17:Bl-l, 41, 43-78
Car#48:Bl-7, 9-40
Car #49: Bl-5, 41-74
Car#50:B6-108
Car #99: 107-1 10 16
There were two express elevators (#6 and #7) to Windows on the World (and related conference rooms
and banquet facilities) in WTC 1 and two to the observation deck in WTC 2. There were five local
elevators in each building: three that brought people from the subterranean levels to the lobby, one that
ran between floors 106 and 110, and one that ran between floors 43 and 44, serving the cafeteria from the
skylobby. All elevators had been upgraded to incorporate firefighter emergency operation requirements.
Local Law 5 (New York 1973) requires that elevators be provided with an emergency recall system. This
requirement was incorporated subsequently into the American Society of Mechanical Engineers (ASME)
A17.1, Safety Code for Elevators and Escalators, which governs elevator design and operation in all
present U.S. building codes. The ASME Code required that:
• All passenger elevators be marked with signs stating that they cannot be used during a fire;
• Fire detectors installed in every elevator lobby and machine room be arranged to initiate a recall
of the elevators to the ground floor where the doors open and the elevator is taken out of service;
and
• Fire service personnel can use a special key to operate any individual car in a manual mode as
long as they feel it is safe to do so.
• At least one elevator serving every floor be connected to emergency power.
Currently, there are no national model codes that permit elevators to be used as a means of occupant
egress in emergencies, and national standard ASME A17.1 (ASME 2000) requires signs at all elevators
warning that they should not be used in fires. There are some recent exceptions to this requirement, but
these are limited to special cases. For example, NFPA 5000 permits protected elevators as a secondary
means of egress for air traffic control towers, and the City of Las Vegas accepted elevators as a primary
means of occupant egress from Stratosphere Tower based on a performance-based design (Bukowski
2003).
16 PANYNJ WTC Fire Safety Plan (Towers One and Two), 1995, WTCI-13-NYC.
34 NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
U.S. building codes (including NYC Building Code) require accessible elevators as part of a means of
egress that may be used by the fire service to evacuate people with disabilities. These elevators must
comply with the emergency operation requirements of ASME A17.1 (Phase II emergency operation by
the fire service), be provided with emergency power, be accessible from an area of refuge or a horizontal
exit (unless the building is fully sprinklered), and operate in a smoke protected hoistway. Phase II
operation involves the use of an elevator by a firefighter for fire service access or for rescue of people
with disabilities performed under manual control (with the use of a special key).
In the event of a fire in WTC 1 or WTC 2, or other emergency requiring evacuation where the stairwells
are unusable or cut off by fire and/or smoke, consideration of using elevators for occupant egress may be
given in accordance with the following PANYNJ guidelines:
• Elevators may not be used if they also service the fire floor, except under specific instructions
from the fire safety director or Fire Department;
• If the elevators do not service the fire floor and their shafts have no opening to the fire floor, they
may be used at the direction of the fire safety director or fire department;
• Elevators under the direction of the fire department or trained building personnel may be used. 7
Every elevator lobby contained a sign reading, "IN CASE OF FIRE USE STAIRS UNLESS
OTHERWISE INSTRUCTED." The sign also included a diagram indicating the location of the sign and
the location and letter designation of each stairwell serving the particular floor. 18
2.2.4 Emergency Communication System
WTC emergency procedures specified that all building-wide announcements were to be broadcast from
the fire command station of each WTC tower, in coordination with the fire safety director or life safety
and security supervisor. The deputy fire safety director was likely to make all announcements.
Appendix J of the World Trade Center Emergency Guidelines 19 provided prepared text for a variety of
emergency scenarios, including power failures, fires, and service interruptions. Prior to all emergency
announcements, the following pre-announcement was made:
"Your attention please, your attention please. An important public
address announcement will be made in the main corridor of your floor in
a few moments."
Evacuation for any reason, including fire or smoke, would have generated the following announcement,
enabling a phased evacuation:
"Your attention please. We are experiencing a smoke condition in the
vicinity of your floor. Building personnel have been dispatched to the
scene and the situation is being addressed. However, for precautionary
reasons, we are conducting an orderly evacuation of floors .
17 World Trade Center Emergency Procedures Manual 2001 - Confidential. Port Authority of New York and New Jersey.
18 The Port Authority of NY & NJ World Trade Center Fire Safety Plan (Towers One and Two). 1995. WTCI-13-NYC.
World Trade Center Emergency Procedures Manual 2001 - Confidential. Port Authority of New York and New Jersey.
NIST NCSTAR 1-7, WTC Investigation 35
Chapter 2
Please wait until we announce your floor number over the public address
system. Then follow the instructions of your fire safety team. We will
continue to keep you advised. We apologize for the inconvenience and
we thank you for your cooperation."
The standard evacuation announcement for a particular floor was:
"Your attention please. It is now time for your floor to be evacuated. In
accordance with the directions from your fire safety team, please take the
exit stairs nearest to your location. We remind you that communications,
emergency lighting and other essential services are in service. We will
continue to keep you advised. We apologize for the inconvenience and
we thank you for your cooperation."
According to the Guidelines, however, the information and instructions broadcast to the building
occupants could be modified to suit the nature of the emergency, at the discretion of the fire safety
director. 20 NIST NCSTAR 1-4 addresses the fire alarm systems in WTC 1 and WTC 2.
Fire Command Station
The fire command station, located in the lobbies of both WTC 1 and WTC 2, provided a command post
for building personnel to orchestrate the response. The NYC Building Code requires that the computer
screen in the fire command station monitor and display information regarding:
Manual fire alarms
Smoke detection
Sprinkler water flow
Elevator lobby smoke detectors
Fire signal activation
Central office notification
Fan system status
Fail safe locked door status
20
World Trade Center Emergency Procedures Manual 2001 - Confidential. Port Authority of New York and New Jersey.
36 NIST NCSTAR 1 - 7, WTC Investigation
Design of the World Trade Center Egress System
Fire system trouble
Fire signal trouble
Tamper switch alarm
Power source
Test/normal mode
Other information as desired,
including the status of
elevators.
The primary value of the fire
command station was its role as a
convening point for key building
personnel responding to a building
incident. The roles of many of the
key personnel are described in the
following section. Figure 2-15 shows
the fire command station in the lobby of WTC 1 on September 11, 2001, seen from the east end of the
Mezzanine Level. The fire command station appears in the back right corner of the picture.
Figure 2-15. Fire command station in lobby of WTC on
September 11, 2001, as seen from mezzanine
escalator, looking west.
2.3
THE HUMAN ELEMENT
PANYNJ produced and regularly updated an emergency procedures manual for building personnel to
follow in the event of a building incident, 21 at least until Silverstein Properties formally had become
leaseholder several months prior to September 11, 2001. While Silverstein Properties was formally
managing WTC 1 and WTC 2, PANYNJ staff continued to be significantly involved in property
management during the transition. The latest update to the manual was completed earlier in 2001. Note
that PANYNJ was not responsible for responding to fires or alarms in WTC 3 (Marriott Hotel), WTC 6
(US Customs House), or WTC 7.
The fourteen chapters in the 2001 manual addressed such possibilities as bomb threats, fires, floods, gas
leaks, elevator emergencies, power failures, medical emergencies, chemical and fuel releases, structural
integrity, and political demonstrations, among other potential problems. Aircraft impact was not
specifically addressed. Individual responsibilities for key personnel were enumerated, including
interactions with non-PANYNJ personnel (including, as appropriate, FDNY, NYPD, and others). The
following subsections of Chapter 2.3, are derived primarily from The 2001 WTC Emergency Procedures
Manual, Chapter 7, entitled "Fire Emergencies."
World Trade Center Emergency Procedures Manual 2001 - Confidential. Port Authority of New York and New Jersey.
NISTNCSTAR 1-7, WTC Investigation
37
Chapter 2
2.3.1 Responsibilities of the Fire Safety Director
The fire safety director was a position required by Local Law 5 (New York 1973). Local Law 5 required
all buildings (new or existing) occupied by more than 100 persons above or below the street level, or
more than 500 people in the entire building to have a fire safety director, deputy fire safety director, and
building evacuation supervisor. Each such building is required to have one employee designated as fire
safety director and one or more employees designated as deputy fire safety director, who possess
certificates of fitness from the commissioner qualifying the individual to conduct fire drills, evacuations,
and related training. A certified individual is required to be on duty during normal working hours.
Consistent with Local Law 5, the primary responsibility of a fire safety director at the WTC (according to
the formal emergency procedures manual) was overall emergency management for a building incident
(PANYNJ 2001b). The fire safety director reports to the Fire Command Station, or scene, and assumes
the following duties:
Verify that FDNY has been notified and coordinate activities of FDNY and other emergency
response personnel;
Confer with floor wardens of affected floor(s) to determine conditions on the floor and identify
areas to be evacuated, route of evacuation, stairwells available, and potential refuge floors;
Initiate evacuation procedures;
Direct public address announcement(s), as necessary;
Deploy security officers to restrict access to affected and secure areas;
Dispatch "key runs"; 22
Ensure appropriate notifications are initiated;
Maintain a chronological record of the event;
Direct the Operations Control Center (OCC) to arrange for emergency elevator service;
Investigate cause of fire (in coordination with the FDNY Bureau of Fire Investigation, prepares
appropriate reports).
2.3.2 Responsibilities of the Deputy Fire Safety Director
The role of the deputy fire safety director assigned to the Fire Command Station in the lobby was to
execute and direct the fire safety plan, including:
• Notify fire safety director, operations control center, and police desk (WTC 5) of incident;
• Maintain communications with floor wardens and other members of the fire safety team;
A "key run" is a security officer charged with distributing and retrieving master key rings and Fire Department portable radios.
These keys provided access to secure areas for emergency responders or building personnel during an incident.
38 NIST NCSTAR 1 - 7, WTC Investigation
Design of the World Trade Center Egress System
• Assist with crowd control and evacuation, as necessary;
• Request that the elevator starter or OCC contact the elevator(s) to respond to the Lobby and await
Fire Department personnel;
• Complete necessary fire alarm notification forms.
2.3.3 Responsibilities of the Operations Control Center Supervisor
Upon notification of a fire event, the supervisor on duty at the Operations Control Center was to first
ensure that the fire command station and the fire safety director were notified. Next, the supervisor was
to issue a general broadcast of information over all WTC radio channels, monitor all channels and ensure
that radio silence is observed unless directly related to the ongoing incident, arrange for elevator service,
update units with relevant information as necessary, and notify managers of Windows on the World
(WTC 1) and Top of the World (WTC 2) of incident in order to "reduce anxiety to tenants, visitors,
guests, etc. when numerous emergency vehicles respond." The Operations Control Center was located in
the Bl Level of WTC 1 and was a backup Fire Command Center.
2.3.4 Responsibilities of the Operations and Maintenance Management
Building operators and maintenance personnel were mobilized in order to provide emergency response
assistance should the need arise. The duty supervisor established contact with the fire safety director, fire
safety coordinator, or life safety and security supervisor and responded to the fire command station to
assist as required. The operations group supervisor, who may have required self contained breathing
apparatus, was assigned to respond to one floor below the scene of the incident, established
communication with the fire command station using the floor warden telephone, assisted with the
evacuation, and kept in contact with the fire command station.
The supervisor of the mechanical contractors was to dispatch staff to the fire pumps in order to "stand by"
for further instructions, dispatched staff to operate the smoke purge system as requested by the fire safety
director or Fire Department, and dispatched staff to secure sprinkler water shutoff valves.
The supervisor of electrical contractors was assigned to dispatch one contract electrician to one floor
below the affected floor in order to assist should the incident involve electrical closets or fixtures, two
electricians to the nearest sub-station below the affected floor, and a supervisor to the fire command
station. Further, the electrical supervisor was to ensure that staff was standing-by in order to secure
electrical power, if necessary, and that portable electrical power was available, as needed, and played a
significant role in post-incident restoration of smoke detectors and/or alarm panels. In the event of a
major disaster, all staff electricians were to report to the electrical shop/office.
The elevator maintenance contract supervisor was to report to the fire command station in order to assist,
as needed, as well as dispatch elevator mechanics to their appropriate posts to assist, as needed.
Figure 2-16 shows a WTC official (denoted by the vest identifying WTC Officials) attempting to
communicate with elevator occupants in WTC 1 on September 11, 2001, from the fire command station
in the lobby.
NIST NCSTAR 1-7, WTC Investigation 39
Chapter 2
2.3.5
Figure 2-16. Elevator communication panel in the fire
command station of WTC 1, as operated on
September 11, 2001.
Floor Warden System
The WTC Emergency Procedures (PANYNJ 2001b) requires each floor of a high-rise building to
designate a floor warden to coordinate the evacuation of the floor, consistent with NYC Building Code.
Assisting the floor warden were deputy floor wardens and searchers, which constitute a tenant fire safely
team. On multi-tenant floors, each tenant identified a floor warden for their space. Once the order to
evacuate a floor was given, those with building authority had specific responsibilities to insure an orderly
evacuation:
• In the event of an emergency, the floor warden was responsible for ensuring that an alarm was
transmitted by either telephoning the police desk or activating a manual pull station. The floor
warden reported the incident in detail to the Fire Command Station, and relayed instructions to
building occupants.
• The floor warden was responsible for notifying occupants of the floor that there was a fire and
ensure that the occupants executed the fire safely plan (PANYNJ 1995). In an emergency,
searchers would round up employees, and the deputy fire warden would move them into the
corridors and make sure all occupants were accounted for. In the event occupants were reluctant
to evacuate, searchers were not required to force evacuation.
• In coordination with the Fire Safety Director, floor wardens selected the safest stairwell to use on
the basis of the location of the fire, including checking the environment in the stair, and notifying
the fire command station which stairwell was utilized.
2.3.6
Occupant and Tenant Training
23
WTC policy was to conduct fire drills every 6 months, consistent with NYC Local Law 5, or shortly
after move-in for all new tenants in WTC 1 and WTC 2. Written procedures specified a three day
Local Law 5-73. §C19-161.2.a.4.
40
NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
advance notice prior to the drill for tenants, through the floor warden and deputy floor warden. The floor
warden then notified all occupants of the floor.
Immediately prior to the fire drill, the public address system would be used to announce that the drill was
about to occur. Occupant attendance at drills was mandatory, with a small "skeleton staff permitted for
business continuity. An occupant who missed a fire drill as "skeleton staff was required to attend the
next fire drill. The occupants were required to assemble outside a designated stairwell.
During the fire drill training, the fire alarm was sounded. The floor warden, deputy floor warden, and
searchers ensured that occupants gathered in the central hallway, near a stairwell. The fire safety team
then instructed the occupants not to attempt to fight fires, not to use the elevators, to obey all instructions
from the deputy fire safety director, and what phone number to call if there was a problem. The location
of the nearest stairwell was identified and the procedures for phased-evacuation (move three floors below
the fire floor, as instructed by the floor warden and/or deputy fire safety director) (PANYNJ 1996).
The standard instruction to the occupants was to evacuate downward (to three floors below the incident
floor). The training did not explicitly instruct occupants not to evacuate upward or attempt to access the
roof. Stairwells A and C went to the 1 10th floors, but only to serve as egress points to descend from the
1 10th floor or the roof. The 1 10th floor was not a re-entry floor, and thus, occupants without an
authorized badge or a key would have been unable to reach the door that led to the roof. Had the 1 1 0th
floor been accessible, actually reaching the roof would have been prevented by two additional doors, in
accordance with Federal Communication Commission regulations. 24 The first door to access the stairwell
to the roof was protected by an access card reader. Upon opening the first door, the individual would
enter a vestibule where, upon showing ID to a closed-circuit television monitored at the Operations
Control Center (OCC), the door would be electronically unlocked from the OCC. Access to the roof was,
thus, limited to a small number of people certified to enter through a radio frequency hazard awareness
class. 24
Floor wardens, deputy floor wardens, and searchers were required as part of their training, to watch a
video, prepared by PANYNJ. The video entitled "WTC Fire Safety" and provided to NIST by PANYNJ,
reviewed the emergency procedures, building fire safety systems, and the responsibilities of the members
of the fire safety team (PANYNJ 1996).
2.4 CHANGES TO THE EGRESS SYSTEM AFTER THE 1993 BOMBING
The February 26, 1993, World Trade Center bombing precipitated a $250 million 25 repair and life safety
upgrade to the complex, including (PANYNJ undated):
• Radio repeaters on the roof of WTC 5 for Fire Department communications.
• Circulation improvements.
The roof housed critical communications equipment, including broadcast facilities for major television stations, paging
transmitters, FDNY transmitters, numerous mobile transmitters. [Information derived from Port Authority response to formal
NIST question, April 2004]
25 Public Comments to Draft Version of NCSTAR 1-7. PANYNJ. 2005.
NIST NCSTAR 1-7, WTC Investigation 41
Chapter 2
New North (to Vesey St.) and South (to Liberty St.) corridors for faster evacuation from the
Concourse (mall).
Two escalators from the Concourse (mall): one to the plaza at WTC 5 and one up to WTC 4 and
onto Church St.
Photoluminescent paint on handrails, stair treads, and stair centerline.
Multiple power sources for stairwell lighting: 2 normal feeds, back-up generator, and a back-up
from the PATH system; battery backup for every other stairwell fixture (up to 90 minutes).
LED exit signs for extra brightness and visibility through smoke conditions.
Fluorescent signs inside stairwells at all stair reentry doors along with raised porcelain type
Braille.
Fire Command Stations in main lobbies.
Two sealed beam (with battery back-up) elevator lights and bells, in addition to normal lighting.
Upgraded elevator intercom system, monitored at Fire Command Station.
New decentralized Fire Alarm System (Style 7), with three separate data risers to transponders
located every three floors; redundant control panels and electronics; multiple control station
announcement capability.
Fire alarm system powered by normal emergency power, battery back-up, and tertiary power to
equipment.
New modernized Operations Control Center with the capability to monitor all HVAC systems
and elevators.
Elevators modernized to current code, including replacing relay system with microprocessor
based system (only 50 percent complete on September 11, 2001).
Sprinkler installation accelerated to completion, including Concourse.
Fire wardens equipped with flashlights, whistles, hats and special training.
Fire drills conducted in conjunction with the Fire Department.
In addition, PANYNJ purchased evacuation chairs, which were provided free to building tenants upon
request, for use by mobility-impaired occupants during emergency evacuation. Further discussion of the
changes to WTC 1 and WTC 2 subsequent to the 1993 bombing can be found in NIST NCSTAR 1-1.
42 NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
2.5 BUILDING CODE ANALYSIS
NIST NCSTAR 1-1 addresses the building codes relevant to WTC, including provisions for egress system
design. For most buildings constructed in the United States, building codes adopted by local jurisdictions
establish minimum requirements for design and construction. However, because the PANYNJ is an
interstate agency, which was established in 1921 under a clause in the U.S. Constitution, its construction
projects are not required to comply with any local or national model building code. For the design of the
WTC towers, which began in 1962, the Port Authority in May 1963 instructed the architect and engineers
to prepare their designs of WTC 1 and WTC 2 to comply with the NYC Building Code. 26 While not
specifically stated in the 1963 letter to the architect, the 1938 edition of the Code was in effect at that
time. In areas where the Code was not explicit or where technological advances made portions of the
1938 Code obsolete, the Port Authority also directed the architect and engineers to propose designs
"based on acceptable engineering practice." When such situations occurred, the Port Authority required
the architect and engineers to inform the Planning Division of the WTC. The Port Authority established a
special WTC office that reviewed and approved plans and specifications, issued variances, and conducted
inspections during construction instead of the city agencies that would normally perform these duties.
In September 1965, the Port Authority instructed the architect and engineers to revise their designs for
WTC 1 and WTC 2 to comply with the second and third drafts of the NYC Building Code that was under
development and to undertake any design modifications necessary to comply with the new code
provisions. 27 Prior to issuance of this instruction, the Port Authority recognized that the draft version of
the new New York City Building Code had incorporated advanced techniques and that the Port Authority
favored the use of advanced techniques in the design of the WTC towers. 28 By adopting the draft versions
of the new NYC Building Code, WTC 1 and WTC 2 were classified as Type 1-B Construction instead of
Type 1-A Construction (see Sect. 9.1.3 for definition and fire protection requirements of Construction
Type), and several architectural features related to egress were modified in the final design (see Sect. 10.1
of NIST NCSTAR 1-1). This relaxation of code requirements allowed the Port Authority to gain
economic advantage. 29 The new NYC Building Code (NYCBC 1968) was enacted by the City Council
on October 22, 1968, approved by the Mayor on November 6, 1968, and became effective on
December 6, 1968.
2.5.1 Egress in the Building Codes
The ability to evacuate thousands of occupants from buildings as massive as WTC 1 and WTC 2, was a
function of three primary variables: how many stairs, how wide the stairs were, and where the stairs were
located. Each of those three factors, in the context of building code requirements, are evaluated below.
26 Letter dated May 15, 1963 from Malcolm P. Levy (Chief, Planning Division, World Trade Department, PANYNJ) to Minoru
Yamasaki (architect, Minoru Yamasaki & Associates) (See Appendix A of NCTAR 1-1).
27 Letter dated September 29, 1965 from Malcolm P. Levy (Chief, Planning Division, World Trade Department, PANYNJ) to
Minoru Yamasaki (architect, Minoru Yamasaki & Associates) (See Appendix A of NCSTAR 1-1).
28 Memorandum dated June 22, 1965 from John M. Kyle (Chief Engineer, PANYNJ) to Malcolm P. Levy (Chief, Planning
29
Division, World Trade Department, PANYNJ) (See Appendix A of NCSTAR 1-1)
Memorandum dated January 15, 1987 from Lester S. Feld (Chief Structural Engineer, World Trade Department) to Robert J.
Linn (Deputy Director, Physical Facilities, World Trade Department) (See Appendix A of NCSTAR 1-1)
NIST NCSTAR 1-7, WTC Investigation 43
Chapter 2
Building codes largely relate required egress capacity to the size of the area served and the nature of the
use of the area served. WTC 1 and WTC 2 were square buildings (roughly 207 ft by 207 ft, measured
internally), with a gross square footage for each tenant floor approximately 42,850 ft 2 (3,990 m 2 ). The
floor areas in the towers were typically one of two use categories: business or assembly. The distinction
is important for calculating egress requirements as the number of people allowed in a given space would
be significantly fewer if the space is used for office (business) activities, than if the same space were used
for assembly activities (such as a restaurant or meeting space). The width and number of stairwells are
then specified to equal or exceed the number of occupants on a floor.
The size of the 'core' varied significantly throughout WTC 1 and WTC 2. Note that the size of the
structural core (as defined by the location of interior load-bearing columns) did not change significantly
from floor to floor. As discussed earlier in Chapter 2.2.3, however, on any given floor, the core space
used for local elevators was reclaimed for leasable office space on successively higher floors within a
zone. For example, while floors 42 to 48 had a core area of approximately 12,000 ft 2 (1,100 m 2 ),
floor 105 had a core area of 6,800 ft 2 (630 m 2 ), or 57 percent of the core area of floors 42 through 48.
The size of the core was important because some building codes (including the NYC Building Code)
calculate occupant load on a net basis rather than a gross basis. A net basis reduces the square footage of
a floor by an amount equal to the unoccupied space on a floor, such as elevator or machinery shafts, and
common areas such as hallways. The logic of using net as a calculation basis rather than gross is that
there does not need to be egress capacity provided for floor area where no occupants would be located.
As an example, while floors in WTC 1 and WTC 2 had a gross square footage of approximately
42,850 ft 2 (3,990 m 2 ), floor 105 had a net square footage of approximately 36,500 ft 2 (3,400 m 2 ), which
yielded an occupant load of 365 persons. Floors 106 and 107 in WTC 1 were designated an assembly
space for Windows on the World and were discussed previously in Chapter 2.2.1. By comparison, the
occupant load on floors 42 - 48 would have been approximately 313 when calculated using 100 persons
per ft 2 net. Generally, however, the floor with the largest occupant load dictates the overall design of an
egress system.
In addition to local changes in the size of the core space, the stairwells in the WTC towers changed floor
location throughout the building, as well. This meant that the remoteness (or the distance the stairwells
are located apart from one another) of stairwells varied, as well. The greatest separation distance between
any two of the three stairwells, as measured by a walking path measurement (assuming that the building is
fully sprinklered, which WTC 1 and WTC 2 were) determines the stairwell remoteness. At the two
extremes of remoteness found in WTC 1 or WTC 2, floors 83 and higher had Stairwell A and Stairwell B
located about 70 ft (21 m) apart, while on floors 77 - 82, Stairwell A and Stairwell C were located
approximately 175 ft - 200 ft (54 m - 63 m) apart (depending upon the walking path on a particular
floor). Coincidentally, WTC 1 was most heavily damaged on floors in the 90s (where the stairwells were
the closest together) and all three stairwells were destroyed, while WTC 2 was attacked in a region where
the stairwells were the most remote (floors 78 through 82) and one stairwell remained passable. The
angle of the airplane impact, the length-wise orientation of the core, and the presence of elevator
machinery near the passable stairwell may also have been contributing factors to the stairwell
survivability, however. For context, most current codes require that two exits be located a distance apart
no less than one-third of the diagonal distance of the area served (if the floor has full sprinkler protection)
or no less than one-half of the diagonal distance of the area served (if the floor is not fully covered by
sprinkler protection). One-half of the diagonal distance of the area served was 147 ft (45 m) and one-
44 NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
third of the diagonal distance was 98 ft (30 m). Thus, in separate areas within the same building, stairwell
remoteness distances would have been less than that required for sprinklered buildings, as well as greater
than that required for unsprinklered buildings.
As described earlier in Chapter 2.2, WTC 1 and WTC 2 had three stairwells, two 44 in. wide and one
56 in. wide. The 44 in. stairwells were served by doors on each floor measuring 34 in. (1 m), while the
56 in. stairwell was served on each floor by a door measuring 44 in. The NYC Building Code was
selected due to the PANYNJ instruction to architects and engineers to adhere to the NYC Building Code.
IBC and NFPA 5000 and NFPA 101 were selected because they are national model codes. See NIST
NCSTAR 1-1 for further discussion of building codes.
2.5.2 New York City Building Code (1968)
Table 6-2 in §C26-601 required 100 ft 2 per occupant (net) for business occupancies, yielding a nominal
occupant floor load of 365 persons per floor, based upon the largest net square footage (36,500 ft 2 ). In
order to provide sufficient capacity for 365 persons, six and one-half units of exit width (at 60 people per
22 in. unit) would have been required, yielding an allowable floor load of 390 people for business
occupancies. A minimum of two stairwells would have been required for an occupant load less than
500 people [§C26-602 Exits from Floors], each equally sized, as no more than 50 percent of the
occupants can be served by a single exit. Two equally sized stairwells would have been 78 in. wide each.
Three stairwells, two 44 in. wide and one 56 in. wide, would also provide the minimum egress capacity
for business occupancy floors.
Floors 106 and 107 in WTC 1 and floor 107 in WTC 2, having had occupant loads of over 1,000 persons
each, would have required four stairwells to serve each floor. Thus, the 1968 NYC Building Code would
have required theses spaces be served by a minimum of four stairwells (as the occupant load was greater
than 1,000 persons). 30 The number of stairwells was not allowed to decrease in the direction egress
travel, therefore, the entirety of WTC 1 and WTC 2 was required to have four stairwells. In 1995, once
these spaces were considered formally between PANYNJ and NYC DOB, three stairwells were shown to
provide adequate total capacity of 1,170 using the NYC Building Code exit reduction clause in §C27-367
(Fasullo 1995). There was no mention, however, of the requirement for a fourth stairwell in either tower
(which existed and continues to exist as a requirement in §C27-366), nor whether the floor and ceiling
system satisfied the area of refuge requirement for a 2 h fire rating.
Each stairwell would be required to have a door at least 0.9 m (36 in.) wide. The sum of two risers and
one tread depth was required to be not less than 0.61 m (24 in.) nor greater than 0.65 m (25.5 in.). 31
(Thus, the 'standard' 7 in. riser and 1 1 in. tread depth would satisfy this formula, at 25 in. [7 in. + 7 in. +
1 1 in.]). According to Table 6-4 in the NYC Building Code, however, the stair rise, may not exceed
Egress calculation performed in 1995 (contained as figures in Chapter 2.2.1) show a calculated egress capactity of
1,170 persons for both floors 106 and 107 in WTC 1 and a calculated occupant load of 1,013 and 1,030 persons for floor 106
and 107, respectively. Note that there was no mention in this memo of a fourth stairwell.
From the NFPA Life Safety Code Handbook (2003 Edition), this note was made about the formula formerly used by NFPA
and currently used by NYC: "This requirement was deleted because it was based on a 300-year-old French formula in which
an inch was a slightly larger unit of measure than it is today. Moreover, people's feet and stride length - the basis of the
formula - were somewhat smaller at that time. Also, the requirement was originally intended only for stairs of moderate
steepness or pitch."
NIST NCSTAR 1-7, WTC Investigation 45
Chapter 2
0.2 m (7.75 in.) and the tread depth must be greater than 0.24 m (9.5 in.). In 1968, the location of a floor
exit was required to be as remote from the others as practicable.
2.5.3 New York City Building Code (October 2003)
As it pertains to the narrow scope of this egress analysis, the requirements related to the egress system in
2003 would be identical to the requirements of 1968, with one significant exception: stairwell remoteness.
In 1968, the location of a floor exit was required to be as remote from the others as practicable. New
York City Local Law (LL) 16 (1984) imposed a remoteness requirement (not retroactive to an exiting
building such as WTC 1 or WTC 2) of 9 m (30 ft) or one-third the maximum travel distance of the floor
(55 m [180 ft]), whichever is greater, which for WTC 1 and WTC 2 was 55 m (180 ft). Thus, all floors of
WTC 1 and WTC 2 had stairwell separations that exceeded the minimum separation distance requirement
of New York City LL 16 (1984).
2.5.4 International Building Code (2000)
Chapter 10 of the 2000 International Building Code (IBC) require 100 ft 2 per occupant (gross), yielding a
nominal occupant floor load of 429 persons per floor. A minimum of two stairwells would have been
required (for occupant load less than 500 persons [Table 1005.2.1]), each equally sized. As WTC 1 and
WTC 2 were fully sprinklered, Table 1003.2.3 requires a minimum of 0.005 m per occupant (0.2 in. per
occupant) totaling 2.2 m (87 in.) of total stairwell width, or two 1.1m (44 in.) stairwells.
Floors 106 and 107 in WTC 1 and floor 107 in WTC 2, having occupant loads of over 1,000 persons
each, would require four stairwells to serve each floor. The four stairwells would be required to be
maintained to grade, as the number of stairwells shall not decrease in the direction of egress travel.
Additionally, the floor system would be required to have at least a 2 h fire rating. If two areas of refuge
were built on floors 106 and 107 (each area holding at least one stairwell), the IBC would permit four
44 in. stairwells.
Section 1003.3.1 requires that each stairwell have a door at least 0.8 m (32 in.) wide. Section 1004.2.2.1
requires that for fully sprinklered buildings, the stairwell doors be located a distance of no less that one-
third the length of the maximum overall diagonal dimension of the building or area to be served (30 m
[98 ft] for WTC 1 and WTC 2). This requirement was met on some floors, but not all floors, as discussed
previously.
2.5.5 NFPA 5000 and NFPA 101 - Life Safety Code (2003)
The gross square footage for each WTC tower was 42,850 ft 2 (3,990 m 2 ). Table 7.3.1.2 requires 100 ft 2
per occupant (gross) for a business occupancy, yielding a nominal occupant floor load of 429 persons per
floor. A minimum of two stairwells would be required (for occupant load less than 500 persons
[Section 7.4]), each equally sized. Table 7.3.3.1 required a minimum of 0.0076 m per occupant (0. 3 in.
per occupant) totaling 3.3 m (129 in.) of total stairwell width, which may be satisfied by two 1.7 m
(65 in.) stairwells, or three stairwells, sized at 1.1 m (44 in.) each. Section 7.2.1.2 requires that each
stairwell have a door at least 0.8 m (32 in.) wide.
46 NIST NCSTAR 1-7, WTC Investigation
Design of the World Trade Center Egress System
Floors 106 and 107 in WTC 1 and floor 107 in WTC 2, having occupant loads of over 1,000 persons
each, would require four stairwells to serve each floor (Section 7.4.1.2). Four stairwells would be
maintained to grade, as the number of stairwells shall not decrease in the direction of egress travel.
Section 7.5.1.3 addresses the remoteness of stairwells, with the purpose "to minimize the possibility that
more than one has the potential to be blocked by any one fire or other emergency condition."
Additionally, Section 7.5.1.3.3 requires that for fully sprinklered buildings, the stairwell doors be located
a distance of no less that one-third the length of the maximum overall diagonal dimension of the building
or area to be served (30 m (98 ft) for WTC 1 and WTC 2). This requirement was met on some floors, but
not all floors, as discussed previously.
2.5.6 Comparison of Current Code Requirements
Differences in Stairwell Occupant Capacity
The IBC allows a reduction in egress capacity for fully-sprinklered buildings, to 0.005 m (0.2 in.) per
person. Thus, while IBC would require two 44 in. stairwells for 429 occupants, NFPA Life Safety Code
would require two 65 in. stairwells from tenant floors in of dimension similar to WTC 1 or WTC 2.
While NYC Building Code also uses 0. 3 in. per person of required exit width, the calculation is net
square feet, rather than gross square feet, effectively reducing the requirements. However, minimum half-
units of exit width (12 in.) used in the NYC Building Code often force the designer to 'round up' the
calculated egress capacity (from 365 to 390, e.g.). Two stairwells, each 78 in. in width, would be a
minimum allowed by the NYC Building Code, as an alternative to the three stairwells (two 44 in., one
56 in.). Table 2-3 summarizes the results of the calculation of the minimum number of stairwells for an
office occupancy with a service area of 42,850 ft 2 gross (36,500 net) using each of the three building
codes described above.
Table 2-3. Minimum stairwell design for 42,850 ft 2 office plan.
Building Code
International Building
Code (2003)
NFPA Life Safety Code
(2003)
New York City Building
Code (2003)
Number and Width of
Stairwells for a
42,850 ft 2 Office Plan
Two Stairwells
44 in. each
Two Stairwells
65 in. each
Two Stairwells
78 in. each
Net vs. Gross Occupant Load Calculations
The IBC and NFPA model codes both calculate the number of occupants per floor (business occupancy)
based on the gross square footage of the floor divided by 100. NYC Building Code, on the other hand,
calculates the maximum occupant load by subtracting from the gross square footage, shafts, storage
rooms, and stairs. Thus, the WTC had a design occupant load of 390 persons per floor, whereas IBC and
NFPA 101 would have required egress capacity for 428 persons per floor. On the other hand, the egress
system would have to reflect the higher occupant load, as well. WTC 1 and WTC 2 would have required
7.5 (rather than 6.5) units of exit width if the occupant load calculation was on a gross basis rather than a
net basis. Thus, three stairwells would need to have been 1.4 m (56 in.), rather than only Stairwell B, to
accommodate the higher occupant load.
NISTNCSTAR 1-7, WTC Investigation
47
Chapter 2
Areas of Refuge and Egress Capacity
When two stairwells are each located in separated areas of refuge, the capacity of each stairwell may be
doubled. NYC Building Code, IBC, and NFPA 101 each permit doubling of a stairwell's capacity using
this method. As discussed in Section 2.2.1, however, NYC Building Code also allows the capacity to be
tripled when three stairwells are each separated from the other two by fire-rated partitions complying with
requirements for areas of refuge. IBC and NFPA 101 do not allow tripling of stairwell capacity.
Stairwell Remoteness
NYC Building Code calculates the minimum separation distance for stairwells based upon a fraction of
the longest travel distance on a particular floor (one-third, if fully sprinklered and one-half if not fully
sprinklered). IBC and NFPA 101 calculate the minimum separation distance for stairwells based upon a
fraction of the maximum diagonal of the floor or area served (one-third, if fully sprinklered and one-half
if not fully sprinklered).
48 NIST NCSTAR 1-7, WTC Investigation
Chapter 3
First-Person Data Collection and Analysis Methods
3.1 PURPOSE AND SCOPE OF COLLECTION OF FIRST-PERSON DATA
The purpose of first-person data collection was to capture the full range of occupant experiences from
World Trade Center (WTC) 1, 2, and 7 from 8:46:30 a.m., when WTC 1 was attacked, until all survivors
had successfully evacuated. The goal was to capture both common (frequent) evacuation experiences and
unique observations or actions that may have contributed to a greater understanding of the events of
September 11, 2001. Potential respondents included all occupants inside WTC 1 or WTC 2 between
8:46:30 a.m. and 10:28:22 a.m. (when WTC 1 collapsed), building personnel, emergency responders, and
family members who spoke to occupants inside WTC 1 or WTC 2 during the attack.
3.2 METHODS
To best capture both the generic evacuation experience and the unique observations and experiences,
multiple interview methods were selected: face-to-face interviews, telephone interviews of a statistically
representative sample of people, and focus group interviews. Each method contributed a unique strength
to the overall objectives, complemented and contributed to understanding the data collected through the
other methods, and established multiple measures of a variety of phenomena. In addition to interviews,
published media accounts, video, and photographs were collected and analyzed. Each method is
discussed below.
3.2.1 Published First-Person Accounts
National Institute of Standards and Technology (NIST) contracted with the National Fire Protection
Association (NFPA) to collect first-person accounts from newspapers, radio and television programs,
e-mail exchanges, and a variety of websites and to distill them into a searchable database (Fahy and
Proulx 2003). Over a period of 18 months, a total of 745 first-person accounts were collected. These
accounts had been published up to 14 months after the event. Although media accounts do not provide
the rigor of a proper scientific study, they do present important insights into events. The objective of the
analysis of the first-person accounts was to gain insight into the variability of human behavior and
response time displayed during the evacuation, and to use the findings as a guide for additional
investigation. For the NIST investigation, the accounts provided background for development of the
telephone survey instrument and aided in identification of individuals with particularly compelling stories
that were of interest for face-to-face interviews conducted as part of the investigation.
A coding tool was developed for content analysis of the first-person accounts. Data were then entered
into the database. The coding tool had 33 questions such as: "On what floor was the person?," "What was
the first cue of the event?," "Was the person injured?," and "What were the conditions in the stairs?" Not
every account provided answers for all 33 questions, since some accounts lacked certain details, but this is
not unlike the situation of a respondent who did not answer some questions in a survey. Once the
745 first-person accounts were summarized, multiple accounts from the same person were merged into
NIST NCSTAR 1-7, WTC Investigation 49
Chapter 3
one, which provided accounts for 465 distinct individuals. (Some survivors provided multiple accounts
through different sources.) Before any analysis began, the database was further limited to the
435 building occupants who were actually in WTC 1 or WTC 2 on September 11, 2001. The accounts
analyzed were from 435 individuals - 251 occupants of WTC 1 and 184 occupants of WTC 2 -
representing occupants from low, middle, and high regions of both WTC 1 and WTC 2.
The content analysis of first-person accounts has significant limitations. First, the actual questions asked
by the journalists reporting the accounts are not usually contained in the accounts. Second, some details
may have been left unreported; and third, more dramatic stories may have been over-represented.
Consequently, while the results of the published accounts analysis cannot be generalized to the overall
population of the towers of the World Trade Center, they provided valuable input to the NIST
Investigation.
3.2.2 Telephone Surveys
The survey objective of the telephone interview phase of this study called for collecting 800 computer
assisted telephone interviews (CATI) of persons occupying either of the WTC towers (WTC 1 and
WTC 2) at the time of the terrorist attacks on September 11, 2001. The sample size of 800 and allocation
of n=400 to each tower were chosen to maximize the statistical precision of estimates and projections
within each tower. 32 Primary statistical analyses were in the form of tabulations and linear statistics
(e.g., reporting of percentages and average/means). Estimates of percentages from tower-specific survey
data (at n=400) exhibit sampling errors no greater than 2.5 percentage points, and 95 percent confidence
intervals of percentages are no greater than ± 5 percentage points. This level of precision was more than
adequate for examining issues of interest in this investigation. Within WTC 1 and WTC 2, independent
proportionate stratified samples of survivors were drawn. In other words, each occupant of a particular
tower had an equal probability of being selected.
Population and Sampling Frame
The total population of people eligible to participate in a telephone interview consisted of individuals who
were inside WTC 1 or WTC 2 between 8:46:30 a.m. and 10:28:22 a.m. on September 11, 2001, with the
exception of emergency responders (FDNY, NYPD, OEM, FBI. Secret Service, ATF, and others). The
sampling frame (i.e., the list from which the sample was drawn) consisted of the names of occupants from
badge lists for persons authorized to be present in WTC 1 and WTC 2 and was assumed to represent the
entire population of individuals eligible to participate in telephone interviews. All occupants who worked
or regularly visited the World Trade Center were required to provide personal data to PANYNJ in order
to be issued a badge to clear through the security station at the entrance of each tower. The badge lists
were provided to NIST by the Port Authority of New York and New Jersey. The lists provide name, floor
32 Multivariate modeling such as correlation analyses, multiple linear regressions, and path analyses, are also a prominent part of
the survey analyses. Like the tabulations, these analyses were conducted independently by tower. A sample size of n=400 per
tower provides more than ample statistical power for the F tests used to determine the significance of the regression models
(i.e., testing the null hypothesis that the ratio of explained variance to error/residual variance is equal to zero). For instance, in
a multiple regression analysis featuring 20 independent variables, the sample size of 400, and 0.05 level of significance (Type I
error), the power of the F test to detect an r statistic (i.e., proportion of explained variance) of 0.06 is just over 81%. See also
Chapter 9 of Cohen, J., 1988, Statistical Power Analysis for the Behavioral Science, Lawrence Erlbaum Associates, Inc.,
Hillsdale, N.J.
50 NIST NCSTAR 1-7, WTC Investigation
First-Person Data Collection and Analysis Methods
of occupancy, employer, and social security number (the only available means of uniquely identifying
individuals).
Representativeness of Badge List
Confidence in the assumption that the badge list accurately represented the WTC population was
increased by comparing independent 'lists' of occupants to the badge list: survivors who were interviewed
by the media and lists of decedents. The three sources of data that were compared were (1) the list of
decedents from CNN web site, (2) a media list of survivors, and (3) the badge list of occupants.
One limitation is that the independent list of media interviewees may not have sufficient information to
indicate whether the listed person should have had a badge (and thus been listed on the badge list).
After comparing the media list with the badge list, it was determined that 134 (93 percent) individuals of
144 selected from the media list were authorized to be at WTC 1 or WTC 2 the day of the tragedy.
Approximately 2 percent of all individuals were definitely not supposed to be at WTC 1 or WTC 2 the
day of the tragedy and insufficient information existed to determine positively if the remaining 5% were
supposed to be at WTC 1 or WTC 2 on September 1 1, 2001.
Based on Table 3-1, a conservative estimate of coverage was obtained by taking the number of persons
from the media list who were definitely authorized to be at WTC (134), and comparing them to the total
possible number of authorized individuals from that list -(134 + 7) = 141. The resulting conservative
coverage rate of the badge list, estimated based upon media interviews with survivors, was (134/141), or
95 percent.
Table 3-1. Comparison of media interviewees and
Dadge list.
Status
Frequency
Percent
Definitely authorized to be at WTC 1 or WTC 2
134
93%
Definitely not authorized to be at WTC 1 or WTC 2
3
2%
Not enough information to determine
7
5%
Total
144
100%
A similar analysis was conducted using the victim list published on the web site of CNN. As Table 3-2
shows, a conservative estimate of coverage was obtained by taking the number of persons from the
decedent list who were definitely authorized to be at WTC (2,141), and comparing that number to the
total possible number of authorized individuals from that list (2,141 + 79 = 2,220). The resulting
conservative coverage rate of the badge list was, as estimated from the CNN victim list, was
(2,141/2,220), or 96.4 percent.
Thus, the assumption that the badge list was a complete universe of possible WTC survivors from which
to select a representative sample was determined to be valid.
NIST NCSTAR 1-7, WTC Investigation 5 1
Chapter 3
Table 3-2. Comparison of CNN victim list and badge list.
WTC Status
Badge List Status
Appears in Badge List
Does Not Appear
in the List
Total
Person authorized to be at WTC
2,141
2,141
Not authorized to be at WTC
N/A
408
408
Insufficient information to
determine WTC authorization
N/A
79
79
Total
2,141
487
2,628
Telephone Interview Sample Selection
The badge list contained September 11, 2001, occupants, occupants who were absent on the day of the
attacks, decedents, former occupants, and non-person listings (false names used in sample testing input by
PANYNJ prior to delivery to NIST but not removed). This meant that a screening effort was needed to
identify "eligible" badge list members - namely, those who were inside WTC 1 or WTC 2 during the
attacks and survived. Moreover, the absence of telephone numbers for the badge holders on the list
necessitated a tracking/locating effort. The primary tracking mechanism was to search public databases
using commercially available batch matching and web-based search utilities. Consequently, a large
sample was needed to generate the 800 completed interviews.
The number of occupant selections drawn into the sample was contingent on four key design parameters:
• The percentage of individuals from badge listings for whom a working telephone number could
be found (initial estimate: 80 percent tracking success)
• The percentage of badge listings that corresponded to a surviving WTC 1 or WTC 2 occupant on
September 11, 2001 (initial estimate: 14 percent)
• The cooperation rate for screening the occupants (initial estimate: 65 percent)
• The interview response rate among September 11, 2001 survivors (initial estimate: 50 percent).
In planning the CATI survey, a number of design parameters needed to be quantified in order to
determine the number of persons to draw from the badge list. The expected disposition of the sample was
developed using the parameters defined in the preceding paragraph. A total sample of 22,735 persons
from the badge list was needed to generate the desired 800 completed interviews. The expected
disposition by tracking efforts, screening and interviewing are discussed later.
A reserve sample of about 14 percent (or about N=3,265) was added in the event additional respondents
were needed due to unanticipated circumstance (if the eligibility rate was actually lower than anticipated).
This brought the total sample size to 26,000. The reserve was initially held "in reserve" while the main
sample was worked. Working the main sample allowed preliminary estimates of all design parameters to
be monitored so that an informed decision could be made on the necessity of releasing none, some, or all
of the reserve.
52
NIST NCSTAR 1-7, WTC Investigation
First-Person Data Collection and Analysis Methods
The badge list contained different counts of persons from each tower (slightly over 50,000 names for each
tower), yet the sample design called for equal samples to be drawn from the collections of badge holders
in WTC 1 and WTC 2. Thus, a disproportionate design (across tower strata) was employed. Within each
tower, independent proportionate samples were drawn using stratification by floor (within tower),
employer (within floor) and last name (within employer). This served to increase the statistical precision
of the tower-specific samples.
Thus, equal-sized samples of 13,000 selections were drawn from each of WTC 1 and WTC 2 badge lists.
Each tower-specific sample was partitioned into 20 random replicates (comprising 5 percent of the total),
and the reserve sample was determined by the last several random replicates for each tower. It is
important to note that all badge holders from WTC 1 floors 92 and above were omitted from sampling
because there were no survivors from those floors.
Table 3-3 summarizes the final disposition of the CATI sample and the total (locating) sample. The table
is comprised of two sets of rows. The top set pertains to the CATI sample and represents those sample
persons for whom an initial telephone number was identified prior to commencing the CATI survey
operations. The bottom set of rows with the heading "Total Sample Disposition" represents the results of
the locating/tracking effort used to identify usable telephone numbers associated with the sample subjects.
(Recall that only name, SSN, and employer were available; no other contact information was readily
available.)
The bottom set of rows shows that telephone numbers were identified for just over three quarters
(76.7 percent) of the sampled subjects. Moreover, this rate was fairly uniform across towers. The
19,923 individuals with an initial telephone number were then loaded into the CATI sample management
system for calling. Ultimately, all reserve respondents were used in the telephone survey. In the initial
design parameters, it was assumed that 82 percent of the subjects would be locatable. While 76.7 percent
is close, many of the numbers were obsolete (e.g., disconnect, wrong number) and necessitated additional
tracking during CATI operations. Ultimately, by the end of data collection, only half the sample
(49.5 percent) represented confirmed contacts with respondents.
The top set of rows in Table 3-3 presents the final disposition of the sample by tower as well as for the
overall sample. Several statistics in the percentage distribution (rightmost) column are notable. First,
NuStats (under contract to NIST) was unable to contact subjects for half the sample (50.5 percent), due to
failures to answer the phone, answering machines, unusable numbers (e.g., wrong number, disconnected,
business), etc. Most of these unusable telephone numbers represent "unbeatable" subjects - subjects for
whom the initial telephone number was incorrect. It bears reiterating that substantial additional attempts
to locate individuals during CATI operations were conducted using powerful subscription-based web-
based search engines. Unfortunately, little information was available for these individuals.
NIST NCSTAR 1-7, WTC Investigation 53
Chapter 3
Table 3-3. Disposition ol
the CATI sample and the total sample by tower.
CATI Disposition:
WTCl"
WTC 2 a
Total
% Distn
Interview
427
376
803
4.0 %
Partial Interview
47
37
84
0.4 %
9/11 decedent
20
40
60
0.3 %
Other decedent
49
39
88
0.4 %
Not Eligible
3,712
3,752
7,464
37.5 %
Language Barrier
135
129
264
1.3%
Eligible Refused to Interview
138
139
277
1.4%
Other Refusal
224
181
405
2.0 %
Respondent not Interviewed
247
168
415
2.1%
Can't contact/locate Respondent
4,987
5,076
10,063
50.5 %
CATI TOTAL
9,986
9,937
19,923
100.0 %
Total Sample Disposition:
WTCl
WTC 2
Total
% Distn
Found initial telephone #
9,986
9,937
19,923
76.6 %
Unable to find a telephone #
3,014
3,063
6,077
23.4%
SAMPLE TOTAL
13,000
13,000
26,000
100 %
a. Table data are unweighted. Tower location as indicated in the badge list and may differ from reported tower location.
Second, the badge list contained a number of ineligible subjects (37.5 percent) - individuals on the badge
list but not in the building on the morning of September 11,2001. An assessment of eligibility rates
appears later. Third, the badge list included decedent names (0.4 percent) - some from the
September 11, 2001, attack (0.3 percent) and others from causes not necessarily related to September 11,
2001 (e.g., cause unknown, natural causes, 0.1 percent). Most of the September 11, 2001, decedent
names were encountered due to a difference between the full (formal) name of the subject and the name
that appeared on the badge list (e.g., the badge list sometimes contained maiden names, middle names,
nicknames, misspelled first or last names, out of sequence names, titles, and so on). This impeded the
ability to remove known decedent names prior to calling.
The final outcome rates of the CATI operations are presented by tower in Table 3-4. The table shows
screening rates, interview rates, and rates of eligible occupants (among those who responded to the
screening questions). The first row shows that screening response rates were relatively uniform across
the towers, at about 46 percent. In other words, approximately 46 percent of successful telephone
contacts resulted in determining whether the potential respondent was present at WTC 1 or WTC 2 on
September 11, 2001. Similarly, interview response rates (among screened eligible subjects) were
relatively stable across towers, at about 49 percent.
The eligibility rates were higher than expected - about 18 percent overall compared to the 14 percent
expected. The eligibility rate among WTC 1 subjects was slightly higher than those of WTC 2.
However, the overall response rates are essentially uniform across towers, at 22.6 percent.
54
NIST NCSTAR 1-7, WTC Investigation
First-Person Data Collection and Analysis Methods
Table 3-
-4. Summary disposition rates by tower.
Disposition Rate"
WTC 1
WTC 2
Total
Screen
46.5 %
45.8 %
46.1 %
Interview
48.6 %
49.5 %
49.0 %
Eligibility
18.9%
16.7%
17.8%
Overall
22.6 %
22.7 %
22.6 %
a. Definitions for "Rates" consistent with American Association of Public Opinion Research (AAPOR) Standards,
which may be found at http://www.aapor.org/pdfs/standarddefs_3. 1 .pdf .
The telephone interview protocol resulted in 803 interviews with individuals who evacuated WTC 1 or
WTC 2 on September 11, 2001, after 8:46:30 a.m. These 803 individuals were interviewed in roughly
equal proportion (N = 440 for WTC 1 and N = 363 for WTC 2) between the two buildings. The interview
results can be generalized to the entire population of survivors in both buildings with a high degree of
statistical confidence.
Telephone Questionnaire
The telephone interview was conducted by trained interviewers using a computer program that provides
questions and answer categories for the interviewer. Prior to being contacted by telephone, subjects
received a letter that outlined the scope and purpose of the investigation and the purpose of the interview,
and indicated that a telephone call would come several days later. A full informed consent statement also
appeared in the letter, as well as in the script for the calls.
When interviewers reached the subjects by telephone, they described the survey, the confidentiality of
responses, the length of the interview, and the voluntary nature of participation. Subjects were then asked
if they wished to participate, which served as the means of obtaining oral informed consent.
The telephone interview instrument (see Appendix A in this report for the complete instrument) included
the questions, variable names, response options, and skip patterns taken directly from the computer
program used by the interviewers. Variable names are used as shorthand for subsequent data analysis.
Questions had a variety of response option categories: multiple choice, interval, Likert scale, or open-
ended. Open-ended responses were minimized where possible due to the analysis burden and the fact that
face-to-face interviews also were being conducted. Skip patterns reduced the burden on the respondent
by skipping questions that would not apply to a particular respondent. For example, a respondent would
not be further questioned about fire drills if he or she had not received fire drill training. Subsequent
discussions of the questions will indicate whether a respondent was read a list of choices or was expected
to give a free response.
The interview, which typically lasted approximately 20 minutes, was designed around five primary
groups of questions, covering emergency training and preparedness, three stages of evacuation
experience, and background information about each respondent.
NISTNCSTAR 1-7, WTC Investigation
55
Chapter 3
Emergency Training and Preparedness
The first group of questions measured the extent to which an occupant had any special level of knowledge
about the building, other than what would be obtained by performing his or her job. The most prevalent
special knowledge would be formal evacuation training, or fire drills. If respondents indicated that they
participated in evacuation training during the 12 months prior to September 11, 2001, further questions
were asked about the content of the training. The occupant's understanding of the emergency procedures,
or the way it was 'supposed to go,' was also measured. Next, a Likert scale 33 measured the perceived
usefulness of the evacuation training in the context of egress experiences on September 11, 2001, ranging
from very helpful to very unhelpful. Finally, the respondent was asked whether he or she knew that there
was a floor warden for his or her floor.
Initial Experience on September 11, 2001
The second group of questions covered the first moments of the September 11, 2001 attack on the World
Trade Center as experienced by the respondent, also known as the initial awareness period. The manner
in which a person first became aware that something was not normal (whether in the building or the
neighboring building) may have influenced subsequent decisions. Examples of awareness channels may
include sensory perception (such as feeling, hearing, or seeing the building shake; seeing or smelling fire
or smoke) or may include a conversation with a person inside or outside the WTC complex. Next, the
respondent was asked to provide context to the initial moment of awareness. Context was first created by
identifying what activity the respondent was performing. Activities included, but were not limited to,
working, conversing with coworker(s), eating, or participating in a meeting. The respondent was then
asked to recall the number of other people he or she was with at the first moment of awareness. People in
groups often defer to group decisions rather than making their own evacuation decisions. Next, a list of
observations was read aloud, and the respondent was asked to indicate whether he or she noticed that
event during the period of initial awareness. These events included smoke, fire, fireballs, collapsed walls,
jet fuel, severely or fatally injured people, sprinklers going on, fire alarm sounding, power outage or
flickering lights, fallen ceiling tiles, and extreme heat. The event proximity was probed for every
affirmative response to determine whether the observed event was in the immediate area or outside the
building. If no affirmative responses were indicated, the respondents were asked whether they observed
any disaster related events not previously mentioned. Finally, the extent of any injuries to the respondent
or those in the immediate area was ascertained, as well as whether the respondent felt that his or her life
or the lives of other people were in danger.
Interim Experience on September 11, 2001
The format of the interim experience group of questions mirrored the format of the initial awareness
questions. The interim time period was defined as the time after initial awareness, but before the person
entered a stairwell or elevator to leave the building. This time period may have ranged from moments to
tens of minutes. The objective of the interim period questions was to determine what motivated/forced
people to either immediately evacuate or delay their evacuation by some period of time.
A Likert scale measures the degree to which the respondent agrees or disagrees with a statement. In this case, the scale
measured helpfulness, including very helpful, helpful, unhelpful, and very unhelpful. A neutral response was not included.
56 NIST NCSTAR 1-7, WTC Investigation
First-Person Data Collection and Analysis Methods
Information about the nature of the event obtained during the initial period often forms the basis for
decision-making during the interim period. For example, many people may have found the environmental
cues from the initial awareness period sufficient to initiate an immediate evacuation. Others may have
required additional information in order to feel comfortable leaving the workplace. Occupants could have
obtained information in two ways: passively and actively. Passive information is information received
without seeking it. In other words, the information is received regardless of whether the person feels it is
needed. Active information is information which the respondent actively seeks and considers important
with respect to his or her decision to evacuate. In the interview, respondents were first asked whether
they received any additional information about the event during the interim period. If so, the source
(who), the nature (what), and the channel (how) of the information was probed. Next, additional
information sought by the respondent was probed, including the source, nature, channel, and whether the
process was successful in gathering additional information.
The perception of risk to the respondent's life, as well as the lives of others, was asked in the same way as
during the initial period, to determine whether the sense of risk was increasing or decreasing over time.
The interviewer probed about the activities of other people in the proximity of the respondent, which may
have influenced the respondent's subsequent choices. Whether other people began evacuating prior to the
respondent was specifically asked. Next, respondents were asked about the activities undertaken during
the interim period, as well as activities that they wanted to carry out but could not. These activities
included work-related actions, such as saving files or shutting machines down; personal actions, such as
gathering belongings or calling people; and emergency-related actions, such as fighting fires/smoke, and
searching for or helping others. If a respondent was unable to accomplish an action, the action and the
reason for being prevented from doing so was captured.
As with the initial period, any observations of building damage were collected. If the respondent received
help in any way before initiating evacuation, the nature and source of the assistance was determined. The
respondent was asked for the primary cue was that initiated his or her evacuation on September 11, 2001
and how many minutes passed before initiating evacuation. Finally, respondents were asked whether
anything prevented them from evacuating sooner than they reported.
Evacuation Experience on September 11, 2001
The next group of questions which followed the evacuation sequence to its completion, focused on time
spent in the stairwell and/or elevator(s). Respondents were first asked whether they began their
evacuation alone or with other people. Which stairwell (or elevator) the respondent entered was collected
as either the stair identification letter (A, B, or C) or the geographic location, if known. Knowing where
the stairwell emptied at the bottom could also narrow down which stairwell was used, which was
collected near the end of this group of questions (Stairs A/C [44 in. wide] emptied out to the upper,
Mezzanine Level, while Stair B [56 in. wide] went to the lower, Concourse Level). Next, the
respondent's rationale for using a particular stairwell was probed. The respondent was then asked
whether he or she left the stairwell or turned back for any reason during the evacuation and, if so, why.
Some events and features of the stairwells aided the progress of the evacuation, while other features
constrained the progress of the evacuation. The following features or events were identified to the
respondents, who were asked to indicate whether it was an aid to their egress: instructions or assistance
from their floor warden, a police officer, or firefighter; support/encouragement from others; exit signage;
NIST NCSTAR 1-7, WTC Investigation 57
Chapter 3
and photoluminscent paint. The following items were identified to determine whether they constrained
the evacuation: crowded stairwells, counterflow (people moving up the stairs, against the flow of
occupants), disabled or injured people being taken down the stairwell, locked doors, poor lighting,
confusing or missing signage, and lack of clear instructions.
As with the initial and interim time periods, environmental cues related to fire, smoke, jet fuel and other
disaster-related observations were probed, as well as whether the observation was in the immediate area
or outside the tower. The final question about the respondent's own evacuation estimated the elapsed
time from entering the stairwell until they left the building. A concluding evacuation question determined
whether they knew why someone on their floor did not survive the WTC attack, if applicable.
Respondent Background
The final group of questions explored the background of the respondent relevant to evacuation. The first
question identified any pre-existing disabilities or injuries which made evacuation more difficult. The
respondent's age, gender, and primary language were collected. If the respondent was working in the
building prior to 1993, they were asked whether they were present during the February 26, 1993 bombing.
If so, respondents were asked questions about their evacuation experience on that day.
The interview concluded with an open-ended opportunity for the respondent to say anything additional
about their evacuation experience on September 11, 2001. Respondents who indicated that they had a
disability, were near the floors of impact, observed fire, smoke, or fireballs in their immediate area, or had
a role of building responsibility on September 11, 2001, were asked if they would be willing to participate
in a follow-up face-to-face interview.
Telephone Interview Response Rate Analysis
The response rate analysis of the telephone interview sample indicated an inverse relationship between
floor height and the rate of response in WTC 1, as shown in the last column of Table 3-5. In other words,
an individual was somewhat less likely to complete a telephone interview if they were high in WTC 1
than if they were lower in WTC 1 . The non-response weight adjustment is the inverse of the overall
response rate. For example, the inverse of 25.3 percent is 3.95. 34 In general, the weight adjustment for
WTC 1 indicates that representative results should reflect that a single interview with a respondent high in
the building is representative of more occupants than a single interview with a person lower in the
building.
1 / 0.253 = 3.95. A lower overall percentage would, therefore, yield a higher weight adjustment.
58 NIST NCSTAR 1 - 7, WTC Investigation
First-Person Data Collection and Analysis Methods
Table 3-5. Response rate analysis for WTC 1.
Floor
Stratum
Number of
Selections
Number of
Interviews
Screen
Eligibility
Interview
Overall
Non-Response
Weight
Adjustment
1 to 42
4464
256
46.2 %
22.6 %
54.8 %
25.3 %
3.95
43 to 75
3714
137
48.6 %
16.6%
45.8 %
22.3 %
4.49
76 to 91
1802
34
42.7 %
14.7 %
30.1 %
12.9%
7.78
Floor missing
6
50.0 %
0.0 %
NA
NA
Total
9986
427
46.5 %
18.9%
48.6 %
22.6 %
Key: NA, not applicable.
A similar analysis of telephone interview response rates for WTC 2 (shown below in Table 3-6) did not
indicate a significant need to weight the results; however, to be consistent with WTC 1 analysis, the
results were weighted.
Table 3-6. Response
rate analysis for WTC 2.
Floor
Stratum
Number of
Selections
Number of
Interviews
Screen
Eligibility
Interview
Overall
Non-Response
Weight
Adjustment
1 to 42
4339
143
44.8 %
14.8 %
49.7 %
22.3 %
4.49
43 to 75
3187
134
45.0 %
17.7%
52.8 %
23.8 %
4.21
76 to 110
2203
94
48.3 %
19.5%
45.2 %
21.8%
4.58
Floor missing
208
5
50.5 %
9.5 %
50.0 %
25.2 %
3.96
Total
9937
376
45.8 %
16.7%
49.5 %
22.7 %
All subsequent telephone interview data analysis in this report reflects weighting of the results in order to
more accurately generalize the results. By convention, when a sample number is indicated (n= ), the
sample number will be the actual number of responses. Where percentages are indicated, however, the
percentages were weighted to allow for generalization, unless otherwise indicated.
The source of differential non-response for floors 76 to 91 in WTC 1 when compared to floors 1 to 75
was not specifically identified.
3.2.3
Face-to-Face Interviews
The objective of the face-to-face interview segment was to gather first-hand accounts and observations of
the activities and events inside the buildings on the morning of September 11, 2001. Using this approach,
NIST identified previously unknown information, evaluated technical hypotheses, and explored conscious
motivations for occupant behaviors, while allowing for comparisons to the telephone interview data.
There was no recording of the face-to-face interviews, other than random selections for quality control
purposes. It is estimated that the average face-to-face interview lasted approximately two hours.
The methodology for the face-to-face interviews is a synthesis of the Behavioral Sequence Interview
Technique (BSIT), originally developed by Keating and Loftus (Keating and Loftus 1984), and the
NISTNCSTAR 1-7, WTC Investigation
59
Chapter 3
Cognitive Interviewing Method (CIM), originally developed by Fisher (Fisher et al. 2000) and Geiselman
(Geiselman 1986). These two interviewing methodologies were developed with the purpose of assisting
persons in retrieving more comprehensive and accurate memories of incidents and sharing important
details. Both approaches begin by allowing the informant to retell an unimpeded account without
interruption by the interviewer, and both initially employ a chronological retelling of information.
However, BSIT was designed to yield a database of qualitative information that could be subjected to
systematic analysis and consolidation, while CIM was designed to facilitate investigative interviews.
Since the Investigation is pursuing both goals (i.e., creation of a database of evacuation-related behaviors
and an investigatory capture of information relevant to outcomes), the methodology combines these two
approaches.
Cognitive interviewing has been the subject of many empirical investigations. Fisher et al. (Fisher,
Brennan, and McCauley 2002) summarized these findings, demonstrating that the methodology
significantly increases the amount of information recalled without affecting rate of errors. Interviewing a
large number of informants will allow corroboration of information, thereby compensating for the likely
increase in the absolute number of errors. Accordingly, it is likely that this approach will be productive in
achieving a holistic view of the building evacuations.
The face-to-face interview methodology involved occupants who may have observed (knowingly or
unknowingly) events important to completion of the objectives of the investigation.
Enumerating the population: The population included the entire occupant and building management
population of World Trade Center WTC 1,2, and 7.
Selecting the sample: The face-to-face interview sample was developed by identifying: (a) individuals
who identified themselves as being willing to share information relevant to the objectives of the NIST
investigation, (b) individuals identified from the telephone interview sample as having experiences or
observation requiring further exploration, and (c) the snowball quota sample approach. A snowball quota
sample approach asks individuals for the names of other people who may meet the selection criteria for
the study. The people identified are subsequently contacted and asked the same question. The process
continues until the quota has been reached.
Data Collection: The face-to-face interviews followed a four step technique, including unimpeded, open-
ended narrative, a structured narrative, technical probes, and closed-ended questions. Each step is
described more fully below.
Step 1: Unimpeded open-ended narrative account: Both BSIT and CIM begin the process by asking
the participant to chronologically recount his or her "story." For the NIST investigation, the starting point
was when it became apparent that something unusual had occurred on the morning of
September 11, 2001. The ending point was when the participant felt that he or she reached a location
where he or she felt safe (or, alternatively, when he or she successfully reached the exterior of the
building). Researchers and practitioners involved with cognitive interviewing believe that starting the
face-to-face interviews in this manner both improves recall and helps build rapport between the
participant and the interviewer. Fisher et al. also noted that asking questions may interfere with recall
because a participant must divide his or her mental resources between recall and listening to the
interviewer's questions (Fisher, Falkner, Trevisan, and McCauley 2000). During the open-ended
60 NIST NCSTAR 1-7, WTC Investigation
First-Person Data Collection and Analysis Methods
narrative account, the interviewer often records notable information that was used for the probing phase
conducted later.
Step 2: Structured narrative account: After participants completed their stories, interviewers prompted
the respondents to go through the story again, but this time working cooperatively with the interviewer to
record entries into a table. This approach was employed by BSIT for three primary reasons: (1) to yield a
structured account that can be entered into a database without further processing; (2) to avoid the biasing
effects of having interviewers ask specific questions; and, (3) to enhance the effort at recall put forward
by participants by encouraging their active collaborative participation, an advantage to open-ended
formats as noted by Fisher et al. (Fisher, Falkner, Trevisan, and McCauley 2000).
Each row of the table represented a single action. The approach was used based on the hypothesis that
people encode narrative memory in a manner consistent with this format, thus facilitating both recall and
data entry. Each column of the table represented three essential components of actions: a cue, an action,
and the reason for taking that action. Cues can be either external (e.g., signs of a fire, someone saying
something) or internal (e.g., remembering another means of escape). Actions are expressed using specific
action verbs (i.e., "ran" instead of "went") and may include objects (e.g., a fire extinguisher) used by the
informant. Reasons are the intentional, goal-directed base for the action. The interviewer used the
participant's own words to the greatest extent possible. Participants were asked to review the data for
accuracy.
Table 3-7 is an example of actions recorded in this manner.
Table 3-7. Example tabular face-to-face interview data entry.
Cue
Action
Reason
I heard but couldn't see
someone yell "I've found a
clear path"
So I stumbled in the dark towards where
I thought the voice came
So that I could find a way to
escape
My path was blocked by debris
So I called out to whoever yelled, "I'm
near the reception area. Where are
you?"
To try to get a better idea about
where the person was
Experimental findings in psychological research on memory (Nillson 2000) suggest that when people
perform actions, their abilities to verbally recall those actions are significantly improved. Script theory
(Schank and Abelson 1977) suggests that people naturally organize their knowledge of actions using
narrative sequences of actions structured around their pursuit of goals. However, in the case of
WTC interviews, gaps in the narrative are anticipated, especially given the long period of time that has
elapsed between the event and the interview. For the NIST investigation, however, interviewers
encouraged participants to report only those memories about which they were confident really occurred to
them.
Step 3: Probing for specific information: After completing the structured narrative account,
interviewers asked specific open-ended questions (probes) intended to elicit specific information of
particular value to the investigation. While some of this information was likely to have been part of the
structured narrative account, participants could also recall other valuable information as well in response
to probes.
NISTNCSTAR 1-7, WTC Investigation
61
Chapter 3
Interviewers may use "context reinstatement" from CIM to improve recall of important information,
because laboratory experiments have demonstrated that contextual cues enhance recall of related
information. Fisher et al. explain that context reinstatement may enhance recall because people use
multisensory coding of events. Using this mnemonic method, interviewers ask participants to "mentally
recreate the external environment, and their affective, physiological, cognitive, and emotional states that
existed at the time of original event" (Fisher, Falkner, Trevisan, and McCauley 2000).
Depending on the population, probes were used to try to elicit information including, but not limited to:
• Location of the informant at the time of certain marker events (e.g., location in WTC 1 when
WTC 2 collapsed);
• Fire conditions (e.g., fire and smoke);
• Other cues of interest (e.g., the smell of jet fuel);
• Presence and activities of persons with disabilities;
• Use of elevators by self or others; and
• Knowledge of any obstacles to their progress while using the stairs.
Because information about many of these areas of concern required precise responses, questions for open-
ended probes were developed collaboratively between the contractor and NIST. Responses to probes
were recorded using standardized formats where feasible. For example, all participants who observed
smoke were asked to estimate the smoke density qualitatively.
Quality Control for Face-to-Face Interviews: With the respondent's consent, some of the face-to-face
interviews were audiotaped. The audio tapes were used for quality control of both data collection and
interview quality. The audiotapes were periodically reviewed to ensure that interviewers precisely
followed the protocol and conformed to administrative requirements.
Interview Responses
Over 200 face-to-face interviews were conducted with survivors from WTC 1 and 2; 131 interviews were
performed with survivors of WTC 1, and 73 interviews with survivors from WTC 2, with the remainder
from WTC 7 or with building personnel who moved between both WTC 1 and WTC 2. An attempt was
made to interview occupants from each zone: low, middle, and high (at least 20 occupants were
interviewed from each zone) of WTC 1 and WTC 2), as well as to interview occupants with unique
experiences: occupants who witnessed fireballs, occupants with mobility impairments, occupants trapped
in elevators, and occupants near or above the floors of impact. Occupants with special roles in the
building, such as floor wardens and PANYNJ employees or contractors, were interviewed. Finally, six
family members who spoke with occupants inside WTC 1 or WTC 2 after 8:46:30 a.m. on
September 11, 2001 were interviewed.
62 NIST NCSTAR 1-7, WTC Investigation
First-Person Data Collection and Analysis Methods
3.2.4 Focus Groups
Williams reported that in a group setting, people provide cues that evoke memories in others and that
social pressures mediate against reporting misrepresentations of what they recall (Williams 1990). The
overall goal of the WTC focus group interviews was to elicit accurate group representations of specific
events or themes. The six focus groups and the corresponding objectives were:
•
•
•
Occupants located near the floors of impact. The objective of conducting a focus group with
people near the floors of impact was to obtain information on the extent of the building damage
and how the damage influenced the evacuation process.
Floor wardens. The objective of the floor warden focus group was to explore the implementation
of the floor warden procedures and the effect those actions had on the evacuation of the occupants
on a floor and the evacuation of the floor warden themselves.
Mobility-impaired occupants. The objective of this focus group was to explore the effect of a
disability on the evacuation of the occupant and any other individuals who may have assisted or
otherwise been affected by the evacuee.
Persons with Building Responsibilities: The objective of the focus group with persons with
building responsibilities was to capture the unique perspective custodians, security, maintenance,
or other building staff.
Random evacuees in WTC 1. The objective of the focus group with random evacuees in WTC 1
was to further explore the variables used in the causal modeling (i.e., those obtained from the
hone interviews) that best explained evacuation delay and normalized stairwell evacuation time,
including environmental cues, floor, and activities.
• Random evacuees in WTC 2. The objective of the focus group with random evacuees in WTC 2
was to further explore the variables used in the causal modeling (i.e., those obtained from the
hone interviews) that best explained evacuation delay, including environmental cues, floor, risk
perception, and use of elevators.
Sample selection: The people selected for inclusion in this study were selected using non-probability
sampling procedures, i.e. snowball quota sampling (Blalock 1972; Cochrane 1977). Respondents
contacted for face-to-face interviews or for other reasons were asked to provide the names and contact
information for people they knew in each of the categories described above. Every effort was made to
include approximately five people in each of these categories in the focus group study.
Data collection: The data collected in the focus group study produced qualitative and detailed narrative
accounts of the experiences of several groups of people, including those near the floors of impact and
those having mobility impairments. The focus group discussion was moderated by a trained and
experienced facilitator. Two notetakers recorded the discussion and later compiled into a single summary
of the focus group.
NIST NCSTAR 1-7, WTC Investigation 63
Chapter 3
3.2.5 Audio, Video, Photographic, and Records Collection
Numerous emergency communications were recorded between 8:46:30 a.m. and 10:28:22 a.m., a period
during which building occupants, WTC personnel, emergency responders, and people outside the
WTC complex used radios and telephones to cope with the unfolding disaster. Emergency 9-1-1 call
records, made available by the City of New York, were analyzed in both audio and transcript format.
Communications at the WTC complex, transcripts of which were made publicly available by PANYNJ,
were analyzed in both audio and transcript format. These communications included radio channels
internal to the WTC complex, such as maintenance, vertical transportation, security, and PAPD. They
also included communications external to the WTC complex, such as NYPD, EMS, PATH, and Newark
airport.
In stark contrast to the number of photographs and video taken of the outside of WTC 1, WTC 2, and
WTC 7, very few video or photographic records from inside WTC 1, WTC 2, or WTC 7 survived. A
significant number of records related to the design and maintenance of the egress and communications
systems, however, were collected and analyzed. In particular, records regarding the identities of the
occupants authorized to enter the WTC complex (referred to as the badge list), architectural drawings,
tenant alteration applications, building upgrades, emergency plans and procedures, and training materials
contributed to a better understanding of the egress system performance on September 11.
Finally, complaints filed with the Occupational Safety and Health Administration (OSHA) were
reviewed. The issues raised by complainants, a combination of surviving occupants and families of
victims, guided the development of interview instruments and identified additional avenues of
investigatory pursuit for emergency preparedness and evacuation system performance.
3.3 ANALYSIS OF FIRST-PERSON DATA
A systematic, comprehensive approach was required in order to consolidate the enormous volume of data
collected or made available. Two primary techniques were utilized in order to capture the full range of
the data collection: quantitative and qualitative analysis. Neither technique would have sufficed by itself,
as the two techniques were highly complementary. In general, the quantitative data analysis was used
with the telephone interview data, while the qualitative data analysis was used with face-to-face
interviews, focus groups, and emergency communications. A number of analysis techniques were
considered, however, the approaches described below were selected to provide the highest quality results
across the three interview methods.
Quantitative Data Analysis
The quantitative data analysis, based on to a telephone interview data set collected according to rigorous
statistical standards, provided the ability to generalize findings and conclusions to the entire population of
WTC 1 and WTC 2 survivors. The results of the 803 telephone interviews were archived and analyzed
using SPSS 12.0.1, 35 a statistical analysis software package. This package provided the ability to apply
weights to the data, compute mean, median, mode, skewness, and other relevant statistical measures,
recode data using expert judgment, and automatically produce reports of subsets.
See Federal Government product disclaimer contained at the beginning of this report.
64 NIST NCSTAR 1 - 7, WTC Investigation
First-Person Data Collection and Analysis Methods
Qualitative Data Analysis
The purpose of the face-to-face interviews and focus groups was to explore the events of
September 11, 2001 at WTC 1, WTC 2, and WTC 7 in an investigatory manner, allowing the individuals
to communicate their experiences, observations, and thoughts outside the constraints of a closed-ended
interview format. While a framework of data collection format and a time frame (the morning of
September 11, 2001) was imposed, respondents had complete freedom to express their experiences in
their own words. The primary value of the qualitative data analysis was two-fold: first, to collect
information on the entire range of observations and experiences, and second, to enhance and provide a
deeper understanding of topics explored and generalized in the telephone interviews. ATLAS. ti 4.1 35 was
used to conduct qualitative data analysis. Each face-to-face interview and focus group was coded for over
130 different types of information, resulting in several thousand individual codings. The codings, along
with a brief explanation of each coding, are included in Appendix B at the end of this report. The
resulting dataset was queried for targeted exploration of various factors of interest.
Protection of Human Subjects
To ensure the protection of the respondents and to comply with the Common Rule for the Protection of
Human Subjects, the protocols and informed consent forms for the conduct of the telephone interviews,
face-to-face interviews, and focus groups were reviewed and approved by an institutional review board
and by NIST. NIST NCSTAR 1-7B summarizes the procedures used in first-person interviews.
Confidentiality of Respondents
NIST noted the importance of maintaining the confidentiality of the respondents. Contractors and
retained experts were bound by contractual obligation to protect the confidentiality of all interview
respondents, whether interviewed by telephone, face-to-face, or focus groups. No identifying information
(name, gender, floor number, job title, etc.) has been included in this report. Individuals may have
previously spoken to the media or other individuals about their experiences and observations, however,
which could diminish NIST's ability to protect an identity. Interviews conducted as part of Project 7 were
distinct from interviews with emergency responders (NIST NCSTAR 1-8 2005), which were conducted in
accordance with a separate agreement between the City of New York and NIST and between the
PANYNJ and NIST.
NIST NCSTAR 1-7, WTC Investigation 65
Chapter 3
This page intentionally left blank.
66 NIST NCSTAR 1-7, WTC Investigation
Chapter 4
September 11, 2001, Before the Attacks
4.1
BUILDING POPULATION ON SEPTEMBER 11, 2001
The total number of people inside WTC 1 and WTC 2 on September 1 1, 2001, is not known precisely, but
it is necessary to make a good estimate of that number in order to provide context for understanding the
evacuation of the buildings. On a typical Tuesday at 8:46:30 a.m., some businesses within the
WTC complex would be largely staffed and operational, with others mostly empty, owing to a later
corporate start time. In addition, September 11, 2001, was both the first day of the new school year for
many children, as well as the date of primary election in New York City. Finally, tourists were not yet
traveling to the observation deck in WTC 2 as it had not yet opened. These factors, among others, may
have acted to limit the number of people who were at the WTC complex on that morning.
The total number of building occupants is equal to the sum of survivors and decedents. Section 9.2
contains an analysis of likely decedent locations at the time or aircraft impact. The response rate analysis
for interviews presented in Section 3.2.2 leads directly to a projection of the number of people present in
WTC 1 (8,900 ± 750) and WTC 2 (8,540 ± 920) on September 1 1, 2001, at the time of the first airplane
impact. Table 4-1 indicates that the populations of WTC 1 and WTC 2 were similar (within statistical
uncertainty intervals) and that 17,580 ± 1,180 individuals were inside the towers at 8:46:30 a.m. These
numbers do not include any airplane passengers or crew, emergency responders, or bystanders. The total
population was rounded to reflect uncertainty in the projection and decedent analysis.
Table 4-1. Occupancy estimates on September 11, 2001, by
tower.
Estimate
WTC 1
WTC 2
Total
Estimated Total Population of
Survivors
7,470
7,940
15,410
Estimated Number of Occupant
Decedents 3
1,462-1,533
630 -701
2,146-2,163
Estimated Total Building
Population
8,960
8,600
17,560
a. See Section 9.2 of this report for analysis of occupant decedent locations in WTC 1 and WTC 2.
While Table 4-1 shows the estimated total population for WTC 1 and WTC 2, Table 4-2 shows the
uncertainty calculations at different levels of statistical confidence. 36 The uncertainty is directly related to
the number of interviews: more interviews completed results in less uncertainty in a projection (i.e., the
number of occupants). Thus, the projection for WTC 1 has less uncertainty than the projection for
WTC 2 (427 interviews and 376 interviews, respectively).
' The standard error expressed in Table 4-2 assumes that the probability distribution is approximately normal (Gaussian). The
standard error then defines the interval over which the actual population of each tower (as opposed to the estimated population)
existed with an approximate level of confidence of 68 percent. The range of population expands as the uncertainty decreases,
as shown by the 95 percent confidence (confidence limit at 5 percent) row in Table 4-2.
NISTNCSTAR 1-7, WTC Investigation
67
Chapter 4
Table 4-2. Occupancy estimate uncertainty.
WTC1
WTC2
Total
Number of Telephone Interviews
427
376
803
Standard Error (p)
1.90%
1.92%
1.36%
Standard Error (Total)
750
920
1,180
Confidence Limits at 5 %
±1,470
±1,790
±2,320
4.2
OCCUPANT CHARACTERISTICS
The survey data indicate that occupants of the WTC towers were twice as likely to be male as female
(65 percent male (n=284) for WTC 1 and 69 percent (n=250) for WTC 2). As shown in Table 4-3 below,
the average age of the occupants was mid-forties, with ages ranging from early twenties to mid-seventies,
although one interviewee indicated that she attended the 90th birthday celebration for a fellow
WTC evacuee (not included in the NIST sample) in 2003. The vast majority of respondents (92 percent
(n=739)) spoke English as their primary language. It should be noted that some telephone contacts ended
with a language barrier and that no interviews were conducted in any language other than English.
Table 4-3. Age for telephone survey respondents in WTC 1 and WTC 2.
WTC1
WTC 2
N b
Valid
439
361
Refuse
1
2
Mean
45
45
Median
46
44
Minimum
22
21
Maximum
73
74
a. Mean and Median values are weighted. N, Minimum, and Maximum are unweighted.
b. The sampling frame (badge list) identified 427 persons in WTC 1. 440 persons responded that they were in WTC 1 at 8:46
a.m. on September 1 1, 2001 during the telephone interviews.
Building occupants become familiar over time with a building, including the location of the emergency
egress components and emergency procedures and protocols. The median residence time of the overall
occupant population could thus be a predictor of how likely it was that an individual received training if
the training was conducted every six months.
Tenant and employee turnover at the WTC was not uncommon. Figure 4-1 shows the reported start dates
for respondents in WTC 1 and WTC 2. In WTC 1, 4 percent (n=18) of the occupants had worked in the
building since 1975. Further, 25 percent (n=l 10) had been working in the building prior to the 1993
bombing, although only 16 percent (n=64) of the WTC 1 respondents were present on February 26, 1993.
For WTC 1, 67 percent (n=287) of the occupants had started working in the building in the last four years
(1998 - 2001). The mean residence time in WTC 1 was over 5.6 years, while the median was 2 years
(half the respondents had been there two years or less, while half the respondents had been there longer
than two years).
68
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, Before the Attacks
Occupant tenure in WTC 2 showed a similar trend. While only one respondent had worked in the
building since 1975, 25 percent (n=91) of the respondents had been working in the building prior to the
1993 bombing (with 16 percent (n=59) present on the day of the bombing). Another 51 percent (n=185)
started working in the building at some point in the four years prior to the 2001 terrorist attack. The mean
residence time in WTC 2 (n=360) was 5.9 years, while the median was 3 years.
i i i i i i i i i i i i i i i i i i i i i i i i i i i
YEAR STARTED AT WTC 1
25.0%-
20.0%-
15.0%-
10.0%-
f]
5.0%-
„_„nnnnnn n nnnnnnnn W
I I I I I I I I I I I I I I I I I I I I I I I I I I
; is i
YEAR STARTED AT WTC 2
Note: Percentages are weighted.
Figure 4-1. Employment start date at WTC.
Overall, 7 percent (n=56) had a formal responsibility or special knowledge about the building. These
respondents included fire safety staff, floor wardens, searchers, building maintenance, and security staff.
Approximately 13 percent (n=105) of the respondents were employed by the PANYNJ, which may not
necessarily imply a special knowledge of the building, as some PANYNJ employees had job duties
related to functions outside the WTC.
Some 6 percent (n=52) reported having a limitation that constrained their ability to evacuate. Both
WTC 1 and WTC 2 had roughly the same fraction of the population who reported a mobility impairment
(n=26 in each building). Six percent of the population of WTC 1 and WTC 2 corresponds to roughly
1,000 people. The reported limitations included chronic illness, recent surgery or injury, obesity, elderly,
heart condition, pregnancy, asthma, and other. Of these conditions, the most prevalent (n=20 of 52) was
recent injury (in particular, severe knee and ankle injuries), followed closely (n=16 of 52) by reports of a
chronic illness (such as cancer, leukemia, arthritis), or use of medications which hindered full mobility or
cognitive ability. Four telephone interview respondents (of 52) reported being pregnant or having asthma,
while three (of 52) reported having asthma- like conditions or indicated that age played a role in their
ability to navigate the egress system. One person reported having a heart condition, while no telephone
interviewee reported being blind, deaf, or requiring the use of a wheelchair. While the last three mobility
impairments were not captured in the 803 telephone interviews, face-to-face interviews and published
media accounts did provide information on people with these disabilities. A small number of respondents
reported more than one mobility impairment.
NISTNCSTAR 1-7, WTC Investigation
69
Chapter 4
4.3
PREVIOUS EXPERIENCE
Of the WTC 1 occupants present on September 11, 2001, 16 percent (n=64) were also present during the
1993 bombing. In WTC 1, 60 percent (n=38) of evacuees who evacuated both in February 1993 and in
September 2001 reported that they evacuated immediately in 1993, while 30 percent (n=20) reported that
they waited to evacuate in 1993, and 9 percent (n=6) did not recall. Most (95 percent (n=53)) who were
able to recall their 1993 evacuation decision felt that they made the right decision, while 5 percent (n=3)
did not believe they made the right decision.
Similarly, 16 percent (n=59) of WTC 2 survivors on September 11, 2001, also evacuated in 1993. In
WTC 2, however, only 75 percent (n=42) felt that they made the right decision in 1993 (compared to
95 percent in WTC 1), possibly due to the fact that many more waited to evacuate in 1993 in WTC 2
(69 percent (n=39)) than did so in WTC 1. Only 31 percent (n=17) of those who reported their decision
evacuated immediately from WTC 2 in 1993. It should be noted that the bomb had a more significant
impact upon WTC 1 in 1993.
4.4
OCCUPANT PREPAREDNESS
Consistent with the NYC Building Code (Local Law 5 §C 19-1 62.2. a.4) and Occupational Safety and
Health Administration (OSHA) regulations (29 CFR 1910.38, "Employee Emergency Plans and Fire
Protection Plans"), 7 the tenants of the World Trade Center were required by the Port Authority to
conduct regular fire drills. Further, they were required to appoint employee floor wardens and searchers.
Overall, 66 percent (n=529) of WTC 1 and WTC 2 occupants reported participation in at least one fire
drill in the twelve months immediately prior to September 11, 2001, and a significant proportion of
occupants had taken part in two or more drills during that time. However, 17 percent (n=139) reported
that they had not participated in any fire drills in the 12 months prior to September 11, 2001, and
17 percent (n=135) did not know. Fire drill participation rates were similar between the two towers, as
shown in Table 4-4.
Table 4-4. WTC fire drills in 12 months prior to September 11, 2001. £
Number of Drills
WTC1
WTC 2
None
18%(n=78)
17%(n=61)
1
13%(n=57)
8 % (n=29)
2
21 % (n=90)
24 % (n=88)
3
1 1 % (n=47)
15%(n=53)
4
10%(n=44)
9 % (n=32)
5-11
7%(n=31)
9 % (n=32)
1 2 or more
3%(n=13)
4%(n=13)
Don't know
18%(n=80)
15%(n=55)
a. Percentages are weighted, n values unweighted.
While the NYC Building Code required fire drills every six months, OSHA regulations require fire drills at least annually. The
PANYNJ required fire drills for all tenants every six months.
70
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, Before the Attacks
One of the goals of fire drill training was to make occupants aware of the location of the emergency exits.
Of respondents who reported participation in a fire drill, 93 percent (n=490) indicated that they had been
instructed about the location of the nearest stairwell. However, of the respondents who reported being
shown a stairwell, 82 percent (n=432) did not enter or use the stairwell during the fire drill. Some
17 percent (n=92) reported that they did use the stairs during a drill, while approximately 1 percent (n=5)
reported not knowing whether they had used the stairs. Overall, more than half (51 percent (n=415)) of
the occupants had never used a stairwell in WTC 1 or WTC 2 prior to September 11, while 48 percent
(n=386) had used a stairwell. Two persons reported not knowing whether they had used the stairs
previously. It should be noted, however, that Local Law 5, adopted in 1973, prohibits occupants from
being required by building management or employers to enter or use a stairwell during a fire drill (1973).
This prohibition may reflect social concerns regarding liability of required stairwell use by occupants and
the economic costs of decreased employee productivity. In contrast, the City of Chicago, Illinois,
(reacting in 2002 to the collapse of WTC) now requires twice-yearly fire drills, which "may conclude
when all participating occupants have fully entered and have begun using designated stairwells." The
City of Los Angeles, CA (Sixth Edition, 2002) in §57.33. 19. C, entitled "Emergency Planning and
Evacuation Requirements for High-rise Buildings," requires that a "minimum of one fire drill annually on
individual floors is mandatory. Total building evacuation is not required."
Another goal of the fire drills was to introduce the floor warden system and evacuation procedures. Most
occupants (82 percent (n=528)) with fire drill training were aware that there was a floor warden for their
floor. Approximately 70 percent (n=557) of all occupants reported that they were aware of the evacuation
procedures. When asked what those evacuation procedures comprised, however, answers varied
significantly, including: wait in hallway for further instructions; do not use elevators, use stairs; meet at a
designated site outside the building for a head count; or proceed down (varied number of) flights of stairs
and wait.
In general, most (66 percent (n=212) in WTC 1 and 60 percent (n=T67) in WTC 2) survivors who
received fire drill training, reported that they found the training to be somewhat or very helpful to their
evacuation experience on the morning of September 11, 2001.
NISTNCSTAR 1-7, WTC Investigation 71
Chapter 4
This page intentionally left blank.
72 NIST NCSTAR 1-7, WTC Investigation
Chapter 5
September 11, 2001, 8:46:30 a.m.
Flight 11 Crashes into WTC 1
"I heard the roar of the plane, looked out of the window [and] saw the
plane halfway in the building. I jumped up and ran out into the hallway
and screamed 'Everybody get out!'" Interview 1000749 (NIST 2004)
5.1
INTRODUCTION
At 8:46:30 a.m. on Tuesday, September 11, 2001, American Airlines Flight 1 1, a hijacked Boeing 767, hit
the north face of WTC 1, as shown in Figure 5-1. This impact resulted in a direct hit on seven floors,
from 93 - 99, with additional damage extending several floors above and below the direct impact area.
As Flight 1 1 crashed into the North Tower, all access to safety for those at or above the impact was
destroyed, including both elevators and stairwells. The fate of over 1,300 occupants located above the
91st floor of WTC 1 was sealed at that instant. This chapter focuses on occupant reaction during the
initial moments after aircraft impact.
As the aircraft struck the building, jet fuel on board ignited. Part of this fuel immediately burned off in
large fireballs that erupted at the impact floors. Remaining fuel flowed across the floors and down
elevator and utility shafts, igniting intense fires throughout portions of the buildings. A fireball killed or
injured several occupants in the Concourse Level lobby. Despite the massive localized damage caused by
the impact, as shown in the computer simulation in Figure 5-2 and Figure 5-3, the structure initially
remained standing.
Figure 5-1. WTC 1 impact, 8:46 a.m.
NISTNCSTAR 1-7, WTC Investigation
73
Chapter 5
Stairwell C
njTUTfriJTrnjTfTLri
U W U LU U 111 U W U
TIES
-V_J LEZ-
Stairwell A
Stairwell B
[^r
T7-T
iiu
LUULUU lSI
lfunruriii
juruuiuu
i nn r*i n
^1
r
WTC1
©
■ - '.' ' *-:<>.'j..--*V.'«.-iIy,<Li
a
n rnnfe'^i^p m runn
lM\
WTO
©
Figure 5-2. Computer simulated impact damage to WTC 1 on floor 95 at 0.7 s after
impact with stairwells superimposed.
74
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 8:46:30 a.m., Flight 11 Crashes Into WTC 1
Floors 93 to 98
Cumulative Damage
Severe Floor Damage
Flreproofing
and partitions ' — '
Floor system . 1
structural damage |_
Floor system
removed '"•
Column Damage
Severed ^)
Heavy Damage ^}
Moderate Damage Q
Light Damage
r. ->■
» o.fi *
4. A »Q A
J. J. J.^ i.
J. JJ. -U. XL XL XL
Floor 93
Floor 94
-
T TTT rj
" r r t r t
-
iiiti
n
"\
"
-
H- -'-'-
* * M '
.
t *-
___ - ~ -.
— * * ±&-i
a & »
- - * Pb 4
" —
„_ — —
_-
ii
»-
■
■"■
""
.
Floor 96
Floor 97
Floor 98
Figure 5-3. Calculated damage to floors 93 through 98 in WTC 1.
5.2
INITIAL OBSERVATIONS IN WTC 1
The New York City 9-1-1 call center was quickly besieged with calls, primarily from citizens outside
WTC 1, reporting that the World Trade Center was on fire. Some callers reported observing a
commercial airliner intentionally hitting the building, and a few quickly concluded it was an act of
terrorism. Other callers reported missiles and bombs, while some called 9-1-1 seeking to find out what
was going on, possibly preventing or delaying people in danger from accessing the 9-1-1 system. For
example, occupants from an upper floor of WTC 1, finding they were unable to connect to 9-1-1, called a
business colleague outside the building, who in turn, was able to contact 9-1-1 and report their plight.
3S
The plane strike was immediately obvious to occupants of WTC 1 throughout the building. Even below
the floors of direct airplane impact, building occupants knew that a significant event had occurred, and
many witnessed significant fire, smoke, or building damage. The majority of survivors in WTC 1 felt the
building move from the airplane impact (63 percent, n=277). Table 5-1 shows a summary of how the
survivors became aware that something was wrong. The reported percentages were roughly constant
throughout the building. For example, 60 percent (n=157) in the upper third of the building (floors 78
and above) felt the building move, 62 percent (n=86) in the middle third (floors 44 to 77) felt the building
move, while 64 percent (n=34) of the respondents in the lower third (up to floor 43) felt the building
move. For some, this first cue was extreme; for others, less so. Still, it was clearly a significant event for
all as the following quotes illustrate.
New York City 9-1-1 Emergency Call Recordings, 2001.
NISTNCSTAR 1-7, WTC Investigation
75
Chapter 5
A survivor from a floor in the 70s felt tremendous movement in the
building: "There were large vibrations, the building shuddering, the floor
shaking violently. The initial explosion was a large cracking sound, and
then boom. The building swayed heavily to the South the first time, and
then a couple other times with decreasing severity." Interview 1000103
(NIST 2004).
Forty floors below the impact, the effect was immediate and significant
as a survivor from a floor in the 50s notes: "There were creaking noises
in the closet. I walked out into the hall and stood there. There was no one
else in the hallway. I heard whooshing noises in the closet. The door
blew open from the closet door, causing my chair to hit the desk"
Interview 1000054 (NIST 2004).
Thoughts of terrorism sprang quickly to the minds of many, particularly
those who had survived the 1993 Bombing: "I felt the building sway. I
knew it was really bad. My co-worker said, 'They did it again.'" Focus
Group #4 (NIST 2004)
Even occupants low in the building knew something major had occurred.
A survivor from a floor in the 20s reported: "We felt the impact. The
building swayed about seven times. Debris was falling down on the
street. We gathered our belongings, I shut off the computer and headed
towards the stairwell" Interview 1000559 (NIST 2004).
Table 5-1. How survivors in WTC 1 became aware something was wrong on
September 11, 2001.
Cue
Percent (n=440)
Felt something (building moving, impact, shaking, swaying)
63%
Heard something (boom, crash, explosion, blast, roar, rumbling)
30%
Other, including saw a plane, smelled jet fuel, fell down/fell off chair,
warned by someone
7%
Source: NIST WTC Telephone Survey Data.
Even at ground level, awareness of the incident was immediate. One survivor who was in the Concourse
of the building recounts this experience:
"I was walking through the mall toward Tower One to get to the elevator.
The lights flickered. I stopped in my tracks and looked around. I saw a
brown cloud coming down the center corridor in the lobby, and I feared
for my safety. The brown cloud had a heavy density and reached from
floor to ceiling. It looked to me like it was both smoke and debris. It first
came from the center corridor, but by the time it reached the revolving
doors (a split second later) it seemed to come from every direction. At
this point, the revolving doors exploded. They seemed to vaporize."
Interview 1000046 (NIST 2004)
76 NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 8:46:30 a.m., Flight 11 Crashes Into WTC 1
Like the Concourse Level, elevator lobbies throughout the building were particularly affected, 39 likely by
excess jet fuel ignited by the crash pouring down the elevator shafts. 40 While only 3 percent (n=l 1) of the
survivors reported seeing fireballs in their immediate area at the time of the airplane impact, the
observations from the face-to-face interviews show the extreme nature of these events:
A survivor from a floor in the 80s: "The entire corridor became an
inferno outside our front door. Smoke began to enter our office. There
was also debris falling. ... The fire on the corridor was at least 10 ft high,
and it ran the . . . good length of the corridor. Then I saw a fireball come
down the elevator shaft and blew the elevator doors. The fireball came
right at me; it was a really bright color." Interview 1000055 (NIST 2004)
A survivor from a floor in the 40s: "I saw the elevator in front me had
flames coming out from it. The elevator was closed but the flames came
from the front where the doors meet and on the sides. They reached
about a foot and a half, with the flames standing from the floor to the
ceiling. I saw a chandelier shaking; it was really moving. The corridor
was dim. I also heard people screaming from the [nearby] floor. I felt the
heat on my face and I thought that my eyebrows were going to get
burned. Black smoke starting filling the corridor, it got really dense
really fast." Interview 1000109 (NIST 2004)
A survivor in the basement: "I saw a big bright orange color coming
through the basement with the smoke ... A fire ball came shooting out of
the basement door." Interview 100760 (NIST 2004)
The elevator lobbies were not the only areas of the building damaged at the time of the airplane impact.
Survivors noticed a range of damage and conditions throughout the building, from lost power to fire and
smoke, to missing walls and floors. Table 5-2 shows observations at the time of first awareness. While
some of the observations involve less severe phenomena (fallen ceiling tiles or flickering lights), others
are more extreme, including collapsed walls, fire, and smoke.
Observation of building damage during this initial awareness period were not as consistent over the height
of the tower as the indications of the airplane impact. Table 5-2 presents a summary of observations
reported during the telephone survey. While damage was more severe near the floors of impact, some
damage was also evident at different locations lower in the building. Survivors provide a range of
observations:
A survivor from a floor in the 90s of WTC 1, just below the impact,
recounts the severe damage on the floor: "In the hallway (from the
bathroom to the elevator), there were no walls left (the wall board was
blown off) and the bathroom seemed to be missing (the walls and the
floor). There was a hole in the wall near the elevator (in the hall) and fire
was coming up onto our floor through that hole." Interview 1000052
(NIST 2004)
39
The majority of face-to-face interview respondents who observed fire inside the building, observed flames at or near the
elevator shafts (NIST Interviews 2004).
See NCSTAR 1-5 for further information about the consumption of jet fuel in WTC 1 and WTC 2.
NIST NCSTAR 1-7, WTC Investigation 11
Chapter 5
A survivor from a floor in the 70s in WTC 1 : "To me everything seemed
normal, all the ceilings were fine, the electricity was fine, and the air
conditioning was also working." Interview 10001 18 (NIST 2004)
A survivor from a floor in the 20s in WTC 1: "I was close to the
windows. The windows were broken and I saw things from the office
were going out the window." Interview 1000064 (NIST 2004)
Table 5-2. Observations of building damage in WTC 1 when occupants first became
aware something was wrong on September 11, 2001.
Observation
Percent (n=440)
Fallen ceiling tiles
17%
Power outage/flickering lights
17%
Smoke
10%
Jet fuel
8%
Fire alarms
8%
Collapsed walls
6%
Other events, including fire, fireballs,
injured people, fire sprinklers going off,
extreme heat, debris
45%
Note: Total does not add up to 100 percent because respondents may have observed more than one event
indicative of damage.
Source: NIST WTC Telephone Survey Data.
While a significant event, not all occupants felt their lives were in danger initially. Of the survivors in
WTC 1, 41 percent felt their life was at risk, and 48 percent felt others lives were at risk, at first
awareness. Only 4 percent of the survivors reported being injured by the attack initially, while only
6 percent reported others being injured.
Most of the survivors were with other occupants when the event occurred. One of every eleven survivors
reported being alone at the time. Sixty-one percent were in a group of 10 persons or less, although two
respondents reported being in a group as large as 400. The average reported group size was 23 persons,
while the median group size was 7 persons. This suggests that a few reports of very large groups
(33 respondents [8 percent] reported being a group of greater than 100 people) skewed the average.
During the initial moments after WTC 1 was attacked, occupants above the 91st floor were trapped. A
few occupants below floor 92 but near the impact region were alive, although trapped as well, some in
their offices, others in elevators. Elevators were rendered inoperable. Occupants observed smoke,
fireballs, damaged walls, fallen ceiling tiles, and injured colleagues on many floors throughout WTC 1.
78
NIST NCSTAR 1-7, WTC Investigation
Chapter 6
September 11, 2001, 8:47 a.m. - 9:02 a.m.
Occupants React to the Attack on WTC 1
"We gathered the group together to figure out what we should do. One
exit was filled with smoke and it was dark; we tried to shut the door to
keep out the smoke. The other exit was of no use, [so] we regrouped and
went toward the main exit (towards the elevator). When the smoke was
building up inside the office, I was more inclined to possibly break open
the window and get some fresh air and wait for help; one or two of the
other people insisted that we start evacuating [recognizing] that we were
in very serious trouble. " Interview 1000137 (NIST 2004)
Between 8:47 a.m. and 9:02 a.m., the time period after World Trade Center (WTC) 1 was attacked but
before WTC 2 was attacked, one of the nation's largest building evacuations and emergency responses
began to unfold. Occupants of WTC 1 , aware that something significant was happening in their building,
were assessing their situation, performing necessary duties, and actively seeking a way out of the
building. With no operational elevators available to the occupants, the three stairwells began to fill, not
only with occupants exiting the building, but also with emergency responders entering the building. Near
and above the floors of aircraft impact, the fire and smoke continued to spread, threatening the lives of the
trapped occupants. Some occupants fell or chose to jump from the building. Building fire safety staff
coordinated the response with incoming personnel from the Fire Department of New York, New York
Police Department, Port Authority Police Department (PAPD), emergency managers from the City of
New York, and Federal agents. At 8:47 a.m., an evacuation order for WTC 1 was broadcast over the
vertical transportation channel (Z). The first radio communication regarding evacuation of WTC 1 and
WTC 2 was issued by a senior PAPD officer at 8:59 a.m. Two minutes later, at 9:01 a.m., an instruction
was issued to evacuate all WTC complex building (PANYNJ 2001a). NIST NCSTAR 1-8 contains a
detailed chronology of all radio communications with respect to evacuation.
Analysis of face-to-face interviews showed that many occupants of WTC 2, if afforded a view to the
north or west, often chose to see for themselves what was happening to WTC 1. Once the significance of
the event was verified, they were forced to decide, first, whether to evacuate or stay in place, and second,
whether to evacuate using the stairs or the elevators. Building fire safety staff were deciding whether to
order an evacuation of the occupants of WTC 2, taking into account the safety of WTC 2 occupants and
what effect an evacuation of WTC 2 may have had on the evacuees from WTC 1 and on the incoming
emergency responders (NIST 2004b).
6.1 CONDITIONS WORSEN IN WTC 1
Within minutes of the aircraft impact, occupants above the 91st floor began to assemble in groups of
various sizes, often taking refuge in offices with access to windows. They also began to reach out for
assistance, calling 9-1-1, family and friends, or colleagues. Electricity and phone service in the region
NIST NCSTAR 1-7, WTC Investigation 79
Chapter 6
41
directly impacted by the airplane was apparently disrupted. Electricity and phone service to floors
above the impact floors was maintained, however, as evidenced by a number of 9-1-1 calls from the 100th
to 106th floors during this time period.
Fire and smoke raced upward. As early as 8:50 a.m., occupants on the 106th floors reported worsening
smoke conditions for about 100 people on that floor, some of whom took refuge in a back office. 42
Somewhat later, at 9:00 a.m., WTC Ch. 9 (PAPD Police Desk) received a call from a Windows on the
World manager, who reported that floor 1 07 was "way too smoky" and most people had retreated to
floor 106 (PANYNJ 2001a). 9-1-1 received reports of hazardous smoke conditions on 103, 104, 105,
and 1 06 within ten minutes of aircraft impact. Each of those four floors reported having more than
100 people trapped on the floor. For some, the conditions remained at least partially tenable during this
time period. Others jumped or fell out of the building within minutes of the aircraft impact. Along with
falling building debris, this created a hazardous situation for emergency responders, evacuees, and
bystanders.
Conditions for occupants below the 91st floor were deteriorating as well. More than one-third of the
survivors reported seeing smoke after the initial airplane impact but before they left their initial floor to
begin their evacuation, up from only 10 percent at first awareness. Eighteen percent of the survivors
encountered the smell of jet fuel, up from 8 percent at first awareness. Table 6-1 shows a summary of
these observations compared to those at first awareness. Observations of nearly all conditions increased as
time progressed.
Table 6-1. Observations of conditions in WTC 1 before beginning evacuation.
Observation
At Awareness
During Interim Period
Smoke
10%
35%
Jet fuel
8%
18%
Fallen ceiling tiles
17%
21%
Power outage/flickering lights
17%
17%
Fire alarms
8%
14%
Collapsed walls
6%
10%
Fire
3%
5%
Other events
45%
48%
Note: Total does not add up to 100 percent because respondents may have observed more than one event indicative of
damage.
Source: NIST WTC Telephone Survey Data.
In WTC 1, a number of people below the impact zone were trapped on their floors, unable to either leave
their offices or reach the stairwells. Prior to 9:02:59 a.m., trapped occupants requested assistance by
calling 9-1-1 from an elevator, from scattered floors in the 10s, 20s, 30s, 40s, 60s, and from most floors in
the 80s. Among those requesting assistance, heavy smoke, wall damage, and occupant injuries were
41 New York City 911 Emergency Call Recordings, 2001.
42 New York City 911 Emergency Call Recordings, 2001.
43 91 1 Emergency Call Records, City of New York, 2001.
80
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1
common reports. Figure 6-1 shows graphically the distribution of observed conditions in WTC 1 after
initial awareness, but before beginning evacuation. The information to develop Figure 6-1 was compiled
from every source available to the NIST investigation, including interviews, published accounts,
transcripts of emergency communication channels, and emergency 9-1-1 calls. Note that the "?" denotes
a floor where there was no information found to record the absence or presence of observations. Further,
the absence of an observation on any floor does not positively exclude the presence of that condition as it
may simply not have been reported.
IIIIIIIIIII
■ ■
I Tr
■ ■
???????????
???????????
•
•
1 -
— L
■
IIIIIII III
A
'
Wa
#5
<?i/
>
fV
%
/-^
1
1
1
1
^
*
s?
*
»s
*s
«
Figure 6-1. Observations of building damage after initial awareness but before
beginning evacuation in WTC 1.
6.1.1
Activities and Information
Occupants of WTC 1 engaged in a variety of activities prior to leaving their floor and beginning their
evacuation, including exchange of information, gathering personal items, helping or searching for others,
and fire fighting. Table 6-2 summarizes activities reported in the telephone survey.
Many respondents used the time prior to beginning their evacuation to gather information about the event
or to call family members. Others helped injured coworkers:
A survivor from a floor in the 90s: "I heard a sound that sounded like a
giant aluminum can being crushed and I felt the building tilt. I tried
calling my company's home office but the line for long distance calls
was not in service. I called home to test the phones and to let my family
know that I was okay. I checked to see if our server was still up. I saw a
man bleeding. I got a first aid kit and succeeded in halting the man's
bleeding. We saw debris and smoke and decided it was time to get out. I
NISTNCSTAR 1-7, WTC Investigation
Chapter 6
got my briefcase, a fire extinguisher, and four diet sodas, exited into the
hallway and went towards stairwell C." Interview 1000052 (NIST 2004)
A survivor from a floor in the 30s: "I used a telephone in the trading
room to call my wife. I wanted to see if she had seen anything on the
news and could tell me what was wrong. I called my wife within
4 minutes of the impact of our building—and I got through okay. She
wasn't there and I left a message. Because we saw the place of impact
and fire coming out of the windows above us in our building, a
co worker and I got our personal belongings and headed calmly to the
stairwell." Interview 1000042 (NIST 2004)
Table 6-2. Activities prior to evacuation reported in telephone survey
by survivors from WTC 1.
Activities Before Evacuation
Percent Reporting the Activity (n=440)
Talked to others
70%
Gathered personal items
46%
Helped others
30%
Searched for others
23%
Talked on telephone
16%
Moved between floors
8%
Fought fire or smoke
6%
Shut down computers
6%
Continued working
3%
Other activities
25%
Note: Total does not add up to 100 percent because respondents may have taken multiple actions.
Source: NIST WTC Telephone Survey Data.
Some occupants, on the other hand, started their evacuation almost immediately:
An occupant from a floor in the 60s in WTC 1 : "It felt like the building
was going to fall over. I grabbed my bag to leave the office floor. I was
not waiting for anyone to tell me what to do." Interview 1000122
(NIST 2004)
An occupant from a floor in the 20s in WTC 1 : "I waited for building to
stop shifting. I began to run straight out the nearest exit out of my office
towards Stairway B. It was the nearest exit from my office and
co-workers were just saying let's go this way." Interview 1000064
(NIST 2004)
While only 1 1 percent of the telephone survey respondents reported being given additional information
about the event during this interim period without consciously seeking the information, 28 percent
reported seeking such information. Table 6-3 shows the types of information received and sought by
occupants. The majority of occupants received or sought information about the event, with a smaller
82
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1
number looking for information on whether to evacuate or remain in the office. For those who sought
additional information, 43 percent were unable to find the information they were seeking.
Table 6-3. Information received and sought prior to beginning evacuation in WTC 1.
Information Type
Information Given (n=50)
Information Sought (n=122)
Information about what had happened
57%
81%
Instructions to leave
28%
17%
Instructions to stay
17%
12%
Other, including information about what to
do and to receive assistance in evacuation,
don't know
13%
13%
Note: Total does not add up to 100 percent because respondents may have observed more than one event indicative of damage.
Source: NIST WTC telephone survey data.
Occupants tried to obtain information through a variety of means -face-to-face conversation; telephone,
television, or radio; e-mail or handheld devices; and from building announcements:
A survivor from a floor in the 70s: "I walked to my desk and spoke on
the phone to find out what happened. I went on the Internet and I was
informed of what happened; also through telephone conversations. I
thought it was necessary to look around. I walked around the floor with
the fire warden; I also stopped, looked, and took some pictures. I was
there in 93, and I wanted to wait for some directions from someone,
through the speaker system, fire alarms, etc." Interview 1000576
(NIST 2004)
A survivor from a floor in the 60s: "During the time in which I was
circulating in the northeast side of the floor, I heard a secretary say,
'Aren't we supposed to wait for an announcement?' I saw other workers
who were standing there talking and trying to assess what to do next. I
went into the hall located between the stairwells A & C. People within
the group helped each other make the correct decisions for evacuation."
Interview 1000639 (NIST 2004)
Twenty-seven percent of survivors felt they began their evacuation before the people around them. Not
surprisingly, survivors' perceptions of risk of death to themselves and to others increased as conditions in
the buildings worsened. Table 6-4 shows a comparison of survivors' perception of risk at the point of
airplane impact and in the interim period before they left their floor to begin their evacuation.
Table 6-4. Survivors perception of risk to self and others after airplane impact and
prior to entering stairwell in WTC 1.
Perception of Risk
Others
Self
At awareness
48%
41%
Interim
63%
52%
Source: NIST WTC Telephone Survey Data.
NISTNCSTAR 1-7, WTC Investigation
83
Chapter 6
Although occupants felt at a heightened risk during this time, many occupants helped others before
beginning their own evacuation; 20 percent of the survivors reported being helped by someone; while
30 percent reported helping others. Table 6-5 summarizes the responses of the survivors who received
assistance.
Table 6-5. Sources of help used by occupants prior to beginning their evacuation
in WTC 1.
Source of Help
Percent (n=87)
Co-worker
48%
Police officer/firefighter
16%
Floor warden
12%
Manager/supervisor
13%
Other/don't know
13%
Stranger
8%
Note: Total does not add up to 100 percent because respondents may have received help from more than one person.
Source: NIST WTC Telephone Survey Data.
Requests for guidance or assistance were not uncommon. At 8:48 a.m., a mere two minutes after the
plane hit WTC 1, an occupant from the 78th floor called the Port Authority Police desk requesting
guidance:
PAPD: Port Authority Police...
Male: Yes, uh, we're on the 78th floor, at Hyundai Securities. Do we need to evacuate or
not?
PAPD: Right away.
Male: Right now?
PAPD: Right away.
Male: Okay, which stairs do we take?
PAPD: Uh, whichever is the easiest one nearest without too much smoke and everything.
Try to get the best way down.
Male: Well, there's. . .the hallways are full. . .full of smoke.
PAPD: Okay. If you could find your way down one. . .
Male: Okay, get out right now, right?
PAPD: Right. Right, exactly.
Male: Okay, bye. (PANYNJ 2001a)
The telephone interviews revealed a variety of reasons which caused occupants decide to evacuate their
floor. Table 6-6 summarizes the single predominant reason given by occupants for beginning evacuation
in WTC 1.
Some left due to observations of building damage or movement, others felt in danger, and still others left
because co-workers left or told them to leave. Less than 1 percent said they left because they heard a fire
alarm:
A survivor from a floor in the 80s: "My boss told me that a plane came
into the building. I was at a cubicle with no window view. I was
84 NIST NCSTAR 1 - 7, WTC Investigation
September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1
screaming and crying, my boss came over to my location. We only had
one door to enter or exit the office. The door was blocked with debris.
We saw the ceiling caving in, but I don't recall any smoke or fire at this
point. We began to dig out pieces of ceiling debris to open the door to
exit." Interview 1000722 (NIST 2004)
A survivor from a floor in the 60s: "The big boss, the treasurer, he stated
"Get out now". I grabbed my pocketbook and started walking towards
the stairwell." Interview 1000834 (NIST 2004)
A survivor from a floor in the 20s: "I saw the floors in the hallway were
twisted. The burning stuff outside the window was getting heavier and I
decided that maybe I should look for a stairwell." Interview 1001667
(NIST 2004)
Table 6-6. Single reason given by survivors to begin their evacuation in WTC 1.
Reason
Percent (n=440)
Building movement
20%
Afraid/felt in danger
20%
Was told to evacuate
14%
Friends/co-workers evacuated
9%
Saw debris
6%
Saw smoke
5%
Other, including saw fire, 1993 experience, saw/heard plane, people
panicking, fire alarm going off
26%
Source: NIST WTC Telephone Survey Data.
6.1.2 Emergency Response at the Fire Command Station, Lobby, WTC 1
Within minutes of the initial impact, personnel from the Port Authority, building security, FDNY, NYPD,
FBI, Secret Service, and representatives of Silverstein Properties (principal leaseholder) were assembling
at the Fire Command Station in the lobby of WTC 1, as shown in Figure 6-2. The deputy fire safety
director (a contractor from O'Conner Security) after assisting a woman injured by glass in the lobby,
quickly began to receive and log calls from floor wardens on floors above the mid-rise area, including
floors above the impact area. As the first calls came in, announcements were made to the affected floors,
indicating, in general terms, 'We have received an alarm downstairs and the alarm is being investigated.
Please stand by.' Any information from the floor wardens about the condition of the floor or injuries was
passed to the Fire Department personnel nearby. As multiple floors were reporting incidents, the deputy
FSD took down the floor numbers on a pad and paper and awaited the arrival of a supervisor. Within ten
minutes, it was determined that the attack was a multiple-floor event. Therefore, consistent with
emergency procedures, building-wide public address system announcements were made informing
occupants to evacuate the building using the stairs and not the elevators. Initially, the evacuation script
NIST NCSTAR 1-7, WTC Investigation 85
Chapter 6
Figure 6-2. Fire Command Station in WTC 1 on September 11, 2001.
was used, but later the
deputy FSD simply
told people over the
public address system
not to use the
elevators because they
were crashing and to
use the stairs to leave
the building (Other
Interview 03
[NIST 2004b]).
Unfortunately, the
individual was
unaware of the
condition of the 22nd
floor, where critical
communications
hardware in the hidden security command center lay in ruins, likely preventing any building-wide public
address announcements from reaching the occupants (PANYNJ 2001a; NIST 2004b). After the fact, a
person familiar with the operation of the building suggested that the fire alarm closet on floor 22
destroyed the riser. NIST NCSTAR 1-4 has a more complete analysis of the fire alarm and public address
system in WTC 1 and WTC 2.
The damage on floor 22 was also reported by several emergency responders (NIST 2004) and was noted
several times in the NIST analysis of the published accounts (Fahy and Proulx, 2003). NIST NCSTAR 1-
8 (Chapter 4.4) contains additional information regarding the status of the 22nd floor command desk, as
well as information flow between occupants, 9-1-1, and emergency responders.
6.1.3
Survivors Begin Their Evacuation
Most, but not all building occupants began their evacuation of the WTC 1 before the WTC 2 was hit.
Ninety-one percent of the survivors in WTC 1 reported beginning their evacuation before Flight 175
struck WTC 2. At this point, nearly all observations of types of building damage had become more
widespread than those at first awareness. Survivors recalled a variety of conditions on the floors as they
left for the stairwells, ranging from significant damage to damage insufficient to deter the occupant from
completing a routine task:
From a floor in the 70s in WTC 1: "As I was leaving it didn't seem as
bad as I thought in the office and I decided that I would just walk all the
way down and reassess the situation and go back to the office if things
were ok." Interview 1000129 (NIST 2004)
From a floor in the 60s in WTC 1: "There was smoke and smell of jet
fuel coming from the stairwell. I covered my nose/mouth with tissue.
This smoke wasn't a lot; not to the extent that [you] could choke. The
smoke was coming from the vents/corners of the stairwell." Interview
1000036 (NIST 2004)
86
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1
From a floor in the 60s in WTC 1 : "The first time I faxed my documents,
they did not go through. I felt safe because I watched previous
documentaries and I was informed that I was safe in my building. I went
back to the fax machine." Interview 1000733 (NIST 2004)
From a floor in the 50s in WTC 1: "I heard a chunk of ceiling fall and a
woman screamed. We all stood and looked at each other and we tried to
figure out what happened. We heard the cable snap in the freight elevator
while we were talking, and the woman yelled "follow my voice". I
followed the woman's voice to find where the stairwell was at to get out."
Interview 1000054 (NIST 2004)
From a floor in the 30s in WTC 1: "It [the stairwell] was the closest one
to our office. I opened the doorway to the staircase... There was a lot of
smoke and there was no one in it. I quickly closed the door." The
occupant went to another stairwell down the hall to leave the floor.
1000009 (NIST 2004)
For consistency in evacuation measurement, time to begin evacuation was defined as the interval from
first awareness to the time the respondent left his or her floor to begin evacuation. 44 On average,
survivors in WTC 1 began their evacuation within 6 min. However, it is important to note that the
statistical distribution of time to initiate evacuation was skewed in the direction of longer delays. In other
words, while the most frequent response for survivors in all three zones in WTC 1 was one minute or less
(referred to as the mode in statistics), and 50 percent of occupants had left their floor within 3-5 minutes
(depending on zone), a few individuals took significantly longer (sometimes longer than 30 minutes) to
start evacuating, thus disproportionately affecting the mean time to start evacuation.
Table 6-7 summarizes the quartile, mode, and average times to start evacuation for survivors in WTC 1.
Note that Table 6-7 separates the occupants into lower, middle, and upper floors based upon the location
of the mechanical floors, which roughly divided the building into thirds. The reported times from the
lower floors were not different than the reported times in the middle floors (tail probability from a log-
transformed t-test comparing the two zones was 0.81). The upper floor evacuation initiation delay times,
however, were statistically significant different when compared to both the middle and lower zones
(significant at approximately the 99 percent confidence level). Thus, occupants nearer the impact area in
WTC 1 delayed their evacuation for a longer period of time than occupants of the other two zones. This
could have been due to the increased frequency of fire, smoke, building damage, and injured occupants on
the upper floors, although that is only one explanation. A further discussion of evacuation initiation delay
time and comparisons across regions of WTC 1 and WTC 2 is contained in Section 10.1.
The time to begin evacuation was defined as the time while on the floor of origin due to the fact that, while many people
decided to leave quickly, they often chose to perform several activities prior to actually entering the stairwell. Thus, time prior
to entering a stairwell (or elevator) was a better measure of evacuation delay than a moment when the occupant 'decided' to
evacuate, which may have been significantly prior to actually starting evacuation.
NIST NCSTAR 1-7, WTC Investigation 87
Chapter 6
Table 6-7. Elapsed time (min) to initiate evacuation for survivors from WTC 1.
Time for Survivors to
Initiate Evacuation"
25%
Initiation
50%
Initiation
75%
Initiation
Mode of
Responses
Average Time
(min)
Lower floors
(Basement - 42)
1
3
5
1
5.7
Middle floors (43-76)
1
3
5
1
4.8
Upper floors (77-91)
2
5
10
1
7.4
a. Time to begin evacuation is the time interval from first awareness to the time the respondent left their floor to begin evacuation.
Source: NIST WTC Telephone Survey Data.
As shown in Table 6-8, survivors below the 92nd floor typically reported choosing the stairwell closest to
them at the time. All three stairwells below the impact region were in use throughout the evacuation.
Some found an appropriate stairwell quickly:
From a floor in the 80s in WTC 1 : "The hallway was free of debris and
well-lit. We (my boss, and co-workers, about ten to fifteen) went back to
staircase C and proceeded to evacuate. The door on the staircase was not
damaged. We entered staircase C. The staircase was well lit and fairly
empty." Interview 1000108 (NIST 2004)
From a floor in the 60s in WTC 1: "The building shook and I thought
something tremendous had struck the building. I looked out the building
to see what had happened. The fire alarm went off. I went to the stairwell
(the exit) to evacuate the floor." Interview 1000025 (NIST 2004)
From a floor in the 50s in WTC 1 : "It [the stairwell] was closest, and we
had been trained in emergencies to only use the stairwells, never the
elevators. We (four of us from my immediate office) exited down
stairwell B." Interview 1000106 (NIST 2004)
For others, finding an appropriate stairwell for evacuation was not always a straightforward process, as a
survivor from a floor in the 30s recounts:
"I opened the doorway to the staircase. There was a lot of smoke and
there was no one in it. I quickly closed the door to not bring smoke into
the floor. The group of people that was with me (about 1 people) started
running back to the office. I began running after my coworkers and
yelling at them to come back to find a different staircase. I was trying to
do the right thing, and they were doing the wrong thing based on the fire
drill training we had. The coworkers weren't listening so I let them go
their own way and I went by myself back out to the hallway to find a
different staircase. I walked down the north-south hallway back past the
original stairwell (the one with the smoke in it) and made a right down
the other hallway because I wanted to go with the crowd. There was a lot
of traffic, so it took a little longer. After a couple of minutes, I went into
the stairwell." Interview 1000009 (NIST 2004)
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1
Table 6-8 shows the percentage of occupants that chose each stairwell for evacuation, as well as their
primary reason for selecting a stairwell.
Table 6-8. Stairwell chosen for evacuation in WTC 1.
Stairwell Used for Evacuation
Percent
Stairwell A
17%
Stairwell B
25%
Stairwell C
19%
Stairwell A or C
10%
Don't know
17%
Other, not applicable, used elevator
12%
Reason for Choosing Stairwell
Closest one
66%
Followed others
17%
Was told to use
12%
Other exits blocked
6%
Other, including don't know, used before, best
conditions, not applicable
18%
Note: Total does not add up to 100 percent because respondents may have given more than one reason for
choosing their stairwell.
Source: NIST WTC Telephone Survey Data.
6.2
INITIAL OBSERVATIONS AND REACTIONS FROM WTC 2
Many of the occupants of WTC 2 quickly became aware that something significant had happened in
WTC 1. Table 6-9 shows how survivors in WTC 2 became aware that something was wrong on
September 11, 2001. Most occupants in WTC 2 heard, saw, or felt the event in WTC 1 (81 percent of the
363 respondents). Others were made aware after a short time by coworkers, telephone, or the news media
coverage.
Table 6-9. How survivors in WTC 2 became aware that something was wrong
on September 11, 2001.
Observation
Percent (n=363)
Heard something (boom, crash, explosion, blast, roar, rumbling)
51%
Saw something (smoke or flames, plane, debris)
19%
Warned by someone around me
13%
Felt something (building moving, impact, shaking, swaying)
11%
Other, including contacted via phone, lights flickered, news media
7%
Source: NIST WTC Telephone Survey Data.
NISTNCSTAR 1-7, WTC Investigation
89
Chapter 6
For occupants near windows which faced north and west throughout WTC 2, what they saw made it
instantly clear that the damage to WTC 1 was severe:
A survivor from a floor in the 90s in WTC 2: "I heard a large noise that
sounded like muffled dynamite and looked out the window which faces
the East - Brooklyn Bridge or into Queens to see if the noise was
connected to anything outside. I saw glittering paper which made no
sense and thought it was part of a promotional event, as if some one was
throwing confetti out of a plane. I headed in the direction of the noise
and saw a gigantic red fire ball at the cubicle diagonally from my desk
and smelled gasoline, which I later learned was jet fuel. I went back to
my office to call my [spouse] to inform [him/her] that I was OK, and that
I was leaving the building." Interview 1000001 (NIST 2004)
A survivor from a floor in the 90s in WTC 2: "I felt a wave of heat; very
high temperature. I went to the window on my floor to find out what was
happening. I saw the flames/fire outside through the window (in Building
One, just about the same floor as mine) and I saw everybody going to the
manager. The manager instructed that, 'people without special
responsibilities should evacuate the building.'" Interview 1000632
(NIST 2004)
A survivor from a floor in the teens of WTC 2: "I heard a loud roaring
sound. I thought it was a window washer falling off its tracks. I walked
into my office and looked out of the window. I saw people running in the
plaza, away from building one. I saw paper and dust floating down from
the sky. I saw the fireball coming from Building One. I thought it was an
explosion of Windows on the World. However, I realized that there were
still several floors above the explosion. Seeing the explosion and the
panic in the plaza triggered my evacuation decision." Interview 1000922
(NIST 2004)
For some in WTC 2, however, the event was not as obvious.
A survivor from the 100s in WTC 2: "A co-worker came to my office
and said "There's a fire in the first building, we recommend that you
leave." I grabbed my bag and packed up my belongings in order to leave
the office. I grabbed my belongings for the reason in which I thought I
was going to go to the gym & then would return to the office." Interview
1000767 (NIST 2004)
Although aware of the event, some occupants of WTC 2 were unsure of appropriate action to take since
the event at this time was limited to WTC 1 .
A survivor from a floor in the 50s in WTC 2: "I heard a loud, horrendous
explosion. I turned and faced the source of the noise. I saw debris flying
through the air outside the window. I saw large objects flying by, then
reams of paper - some that were burning - like confetti. I looked up at
One World Trade Center and saw a gaping hole and smoke. I called my
wife and parents to reassure them, to tell them I was okay. Co-workers
and I discussed what our course of action should be. We stood kind of
90 NIST NCSTAR 1 - 7, WTC Investigation
September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1
paralyzed, undecided as to what we should do. The Bloomberg headline
said a plane had hit WTC 1. We didn't know if we should close the desk
and leave; if we should assume the day was over. We were looking for
guidance." Interview 1000557 (NIST 2004)
An occupant from WTC 2 at 8:49 a.m. also sought guidance from the Port Authority Police:
PAPD: Port Authority Police...
[Caller]: Yeah, this is [Caller], Securities Department from Morgan Stanley.
PAPD: Uh-huh?
[Caller]: Uh, what's the status right now as far as (overlap)
PAPD: We're still checking. Everybody just get out of the building right now.
[Caller]: All right. Have you guys announced an evacuation of Two?
PAPD: We are trying to do that right now.
[Caller]: All right, thank you.
PAPD: All right? We are just advising everybody to get out of the building.
[Caller]: All right, thank you, bye-bye. (PANYNJ 2001a)
At 8:53 a.m., an occupant from WTC 2 called the Port Authority Police Department seeking advice.
Male: Hi, um, I'm on the 95th floor of Two World Trade Center.
PAPD: Yeah, just come on down anyway, sir.
Male: Does that mean walk down the stairs?
PAPD: You'd, be advised, right now at this time.
Male: Should we evacuate all of our people?
PAPD: Yes. Yes. Yes. Everybody.
Male: Okay, thank you. (PANYNJ 2001a)
6.2.1 Activities and Information - WTC 2
Occupants of WTC 2 engaged in a variety of activities prior to leaving their floors and beginning their
evacuation, including exchange of information, gathering personal items, helping or searching for others,
and continuing work activity. Table 6-10 summarizes activities reported in the telephone survey.
A survivor from a floor in the 60s of WTC 2 engaged in a variety of activities before beginning her
evacuation.
"I observed that the secretary was very upset. I went with [the secretary]
and my manager to the lobby area to help care for [the secretary]. People
were leaving the building and I didn't know when I'd be returning to the
building, so I gathered my stuff and I went to the [rest] room. I saw
flying things in the air and everyone was in a commotion to see what was
going on, so I went back to the windows in the northern part of the
building to find out what was going on. I heard the announcement . . .
and I went back to my desk. [There] I made two phone calls to my
mother and my [spouse] to reassure them that I was not affected and that
I was safe. Two of my co-workers came into my office area and stated
that they saw people jumping from Building 1 and that we should leave.
NIST NCSTAR 1-7, WTC Investigation 91
Chapter 6
I gathered my stuff again and prepared to leave." Interview 1000877
(NIST 2004)
Table 6-10. Activities prior to evacuation reported in telephone
surveys by survivors from WTC 2.
Activities before Evacuation
Percent Reporting the Activity (n=363)
Talked to others
75%
Gathered personal items
57%
Helped others
34%
Searched for others
32%
Talked on telephone
16%
Moved between floors
8%
Shut down computers
7%
Continued working
6%
Fought fire or smoke
1%
Other activities
20%
Note: Total does not add up to 100 percent because respondents may have observed more than one
event indicative of damage.
Source: NIST WTC Telephone Survey Data.
Another occupant from a floor in the 100s in WTC 2, who had only worked in the WTC for a couple of
months, took relatively few actions before beginning evacuation.
"I heard a female co-worker who had a window cubicle shouting 'Get
out!' I turned around from my inner office, grabbed my purse, and
walked out my office door. I ran to another co-worker and asked her
what was going on. [The co-worker] didn't know and continued
walking. I decided to follow the co-worker to [figure out] where to go
and to find out what was going on. I followed [the co-worker] to a
stairwell and began to go down the stairs." Interview 1000897
(NIST 2004)
While 2 1 percent of the WTC 2 telephone survey respondents reported being given additional information
about the event during this interim period without actively seeking the information, 1 8 percent reported
actively seeking such information. Table 6-1 1 shows the types of information received and sought by
WTC 2 occupants. The majority of occupants received or sought information about the event, with
smaller number looking for information on whether to evacuate or remain in the office. For those who
sought additional information, 39 percent were unable to find the information they were seeking.
Before WTC 2 was hit, information about the event affected occupants of WTC 2 in different ways. Even
if an occupant heard that something happened to WTC 1 , he or she may have still felt safe in the building.
For instance, an occupant beginning evacuation from a floor in the 50s in WTC 2 took to the stairs early,
only to return to his/her desk.
92
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1
"[After a few floors], I thought to myself, 'Why am I running?' I
remembered that I was the fire warden for my floor." The occupant then
took an elevator back to his/her original floor "to see if anyone else was
there. I thought about the training and what I was supposed to do."
After following the training procedure of calling security, the occupant
"picked up the cell phone and called our [out-of-town] office to let them
know that we were evacuating the building. I got ahold of them and they
told me that a plane hit the building and to get out of there." Interview
1001666 (NIST 2004)
Table 6-11. Information received and sought prior to beginning evacuation in WTC 2.
Information
Information Given (n=77)
Information Sought (n=64)
Information about what had happened
65%
92%
Instructions to leave
26%
17%
Instructions to stay
35%
13%
Other, including information to remain
calm or a choice to evacuate or stay.
10%
5%
Note: Total does not add up to 100 percent because respondents may have received or sought information from more than one
source.
Source: NIST WTC Telephone Survey Data.
The occupant then proceeded to look around the floor for other occupants as well as make a phone call
home to a family member. The occupant took an additional phone call when WTC 2 was hit, which
prompted an immediate evacuation. Interview 1001666 (NIST 2004)
However, another fire warden from a floor in the 40s of WTC 2 used observations from and media
information about WTC 1 to begin evacuation immediately.
After seeing paper flying outside the window and smelling gas, "I ran
across the hall and came across the boss and immediately informed [the
boss] that we should leave. I saw the TV and saw what had happened on
CNN and was informed that a plane hit the building, Tower 1 . I went to
the stairs." Interview 1000867 (NIST 2004)
Twenty-four percent of survivors felt they began their evacuation before the people around them. Not
surprisingly, survivors perception of risk to themselves and to others increased as conditions in the
buildings worsened. Table 6-12 shows a comparison of survivors' perception of risk at the point of
airplane impact and in the interim period before they left their floor to begin their evacuation.
Table 6-12. Survivors' perception of risk upon airplane impact and
prior to beginning their evacuation in WTC 2.
Perception of Risk
Others
Self
At awareness
52%
29%
Interim
67%
42%
Source: NIST WTC Telephone Survey Data.
NISTNCSTAR 1-7, WTC Investigation
93
Chapter 6
Although occupants felt a heightened risk during this time, many occupants helped others before
beginning their own evacuation. Seventeen percent of the WTC 2 survivors reported being helped by
someone and 34 percent reported helping others. Table 6-13 summarizes the responses of the WTC 2
survivors.
Table 6-13. How survivors received help prior to
beginning their evacuation from WTC 2.
Source of Help
Percent (n=60)
Co-worker
56%
Manager/supervisor
15%
Floor warden
12%
Other/don't know
11%
Police officer/firefighter
7%
Stranger
5%
Source: NIST WTC Telephone Survey Data.
Even though the environmental cues were more obvious in WTC 1 than they were in WTC 2, 86 percent
of the survivors in WTC 2 began their evacuation before their building was hit. Occupants gave a variety
of reasons for beginning their evacuation, which are summarized in Table 6-14.
Table 6-14. Single reason given by s
evacuation from Wl
urvivors to begin their
rc2.
Reason
Percent (n=363)
WTC 1 observations
26%
Was told to evacuate
21 %
Afraid/felt in danger
17%
Friends/co-workers evacuated
11 %
1 993 experience (thought was a bomb)
6%
WTC 2 building hit
6%
Other, including jet fuel, information seek,
evacuating the building just felt like the right
thing to do
7%
Source: NIST WTC Telephone Survey Data.
Eighty-six percent to 91 percent of the occupants in WTC 2 began their evacuation before WTC 2 was
hit. 45 Approximately one in four WTC 2 occupants began evacuation because of their observations of
WTC 1 from their office windows, while a similar percentage (21 percent) were told to evacuate.
45 86 percent was calculated by summing the number of people who reported starting their evacuation in less than 16 minutes,
whereas 91 percent of people reported starting their evacuation prior to WTC 2 being attacked. These two questions were
asked independently of one another during the telephone interviews. Note that while the discrepancy may be partially
94
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1
An occupant from a floor in the 90s in WTC 2 left because of
observations of WTC 1 : "[I] saw a gigantic red fire ball at the cubicle
diagonal from my desk and smelled gasoline which I later learned was jet
fuel. I went back to my office to call my [spouse] to tell [the spouse] that
I was OK and that I was leaving the building." Interview 1000001
(NIST 2004)
Another occupant from a floor in the 30s in WTC 2 was told by company
management that the occupant "should begin a self-evacuation, instead
of waiting for the building management to tell us to evacuate. I went
back to my desk to prepare to leave. I decided to follow the group that
was heading for the stairwell." Interview 1000049 (NIST 2004)
An occupant from a floor in the 90s in WTC 2 "heard a noise that
prompted me to get up and look out the window. I saw people out of the
corner of my eye, grabbing their bags and leaving. I turned and got my
bag, with my pocketbook and things to leave my cubicle and follow the
people [to the staircase]. I didn't have to investigate; I just left [because]
I saw other people leaving." Interview 1000070 (NIST 2004)
6.2.2
Survivors Begin Their Evacuation
On average, survivors in WTC 2 began their evacuation within 6 min. However, as with WTC 1 , the
statistical distribution of time to initiate evacuation was skewed toward longer preparation periods.
Overall, the most frequent response for survivors WTC 2 was one minute or less (although the middle
floors had a mode of 5 min), and 50 percent of occupants had left their floor within 3-5 minutes
(depending on zone), a few individuals took significantly longer (more than 30 minutes) to start
evacuating, thus disproportionately affecting the average time to start evacuation. Table 6-15
summarizes the quartile, mode, and average times to start evacuation for survivors in WTC 2. Note that
Table 6-15 separates the occupants into lower, middle, and upper floors, based upon the location of the
mechanical floors which roughly divided the building into thirds. The evacuation delay results may be
biased for the upper floors in WTC 2, however, as only occupants who acted quickly to move below the
78th floor before 9:02:59 a.m. could be interviewed. In other words, those who delayed for whatever
reason, with few exceptions, did not survive. The impact of this potential bias was not quantified, but
should be noted. A further discussion of evacuation initiation delay time and comparisons across regions
of WTC 1 and WTC 2 is contained in Section 10.1.
Table 6-15. Elapsed time (min) to initiate evacuation for survivors from WTC 2.
Time for Survivors to
Enter Stairwell 3
25%
Initiation
50%
Initiation
75%
Initiation
Mode of
Responses
Average Time
(min)
Lower floors
(Basement - 42)
1
4
5
1
6.3
Middle floors (43-76)
2
5
10
5
7.1
Upper floors (77-1 10)
l b
3 b
5 b
l b
4.2 b
a. Time to begin evacuation is the time interval from first awareness to the time the respondent left their floor to begin evacuation.
b. The evacuation delay results may contain significant bias for the upper floors in WTC 2, as discussed in the text.
Source: NIST WTC Telephone Survey Data.
attributable to the uncertainty in time recollection by the occupants, the uncertainty intervals for the two percentages overlap at
the 95 percent confidence level (plus or minus 5 percentage points).
NISTNCSTAR 1-7, WTC Investigation
95
Chapter 6
Occupants typically chose the stairwell closest to them at the time. Below the impact region, all three
stairwells were in use throughout the evacuation. Table 6-16 shows the percentage of occupants each
stairwell for evacuation and why that stairwell was chosen.
Table 6-16. Stairwell chosen for evacuation in WTC 2.
Stairwell Used for Evacuation
Percent
Stairwell A
18%
Stairwell B
18%
Stairwell C
14%
Stairwell A or C
10%
Don't know
14%
Other, including closest to the office, not applicable,
used elevator
27%
Reason for Choosing Stairwell
Closest one
63%
Followed others
20%
Was told to use
10%
Other exits blocked
4%
Other, including don't know, used before, best
conditions, not applicable
13%
Note: Total does not add up to 100 percent because respondents may have given more than one reason for
choosing their stairwell.
Source: NIST WTC Telephone Survey Data.
An occupant in the 80s of WTC 2 notes that colleagues "ran down the stairs in the South part of the office
to evacuate the building. The staircase that I chose was the only one that I was aware of." Interview
1000003 (NIST 2004)
Another occupant in the 60s moved to the stairwell that was already being used.
"I saw a person opening the door to a stairwell to the left of the reception
area, near the men's room. We walked really fast to enter the stairwell to
make our way down." Interview 1000526 (NIST 2004)
An occupant in the 100s of WTC 2 was aware that he/she was assigned to use Stairwell B.
"Due to the fire drill training, we were always told to use Stairwell B. I
began going down Stairwell B in order to avoid elevators and to get out
of the building safely." Interview 1000906 (NIST 2004)
6.2.3
Elevator Use in WTC 2
Elevator usage by occupants played a significant role in reducing the total loss of life in WTC 2 on
September 11, 2001. Sixteen percent of the occupants of WTC 2 used elevators for at least part of their
96
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1
egress from the building; another 2 percent used elevators in the basement levels. Occupants higher in the
building were somewhat more likely to use elevators than occupants lower in the building: subdividing
the 16 percent who used elevators, 4 percent were from floors below the 44th floor skylobby, 5 percent
from floors 46 to 73, and 7 percent from floors above the upper skylobby on the 78th floor. Expressed
another way, 12 percent of survivors in the lower floors reported that they used an elevator, while
15 percent of survivors from the middle floors reported using an elevator, and 33 percent of survivors
from the upper floors reported that they used an elevator to evacuate WTC 2. Interpretation of the latter
percentage may be biased, however, as occupants who used stairwells or delayed evacuation may have
been disproportionately trapped above the impact region after 9:02:59 a.m. when compared to occupants
who initially chose to evacuate using elevators over stairwells.
For some occupants, the physical challenge of navigating the stairwells was daunting enough to
precipitate rapid use of the elevator system.
An occupant from the 90s: "I think perhaps a survival instinct prompted
my decision. I grabbed my briefcase and walked fast toward the
elevator. I saw someone who needed to use crutches just as the elevator
door was closing on the floor. We opened the door for him manually;
the 'open door' button never functioned properly. I was taking a new
medication and knew I should not walk down the stairs, so I took the
elevator to the 78th floor and then took another elevator to the lobby."
Interview 1000553 (NIST 2004)
For others, physical challenges presented themselves while in the stairwells. Even before WTC 2 was
attacked, Port Authority communications (Channel 27) indicate that an occupant required assistance in the
stairwells at 9:00 a.m. (PANYNJ 2001a) Others were able to switch egress modes from the stairwells to
the elevators, although not without some difficulty, as one occupant from the 40s demonstrated:
"My leg was hurting (it was a pre-existing condition) and I had
bronchitis. The security guard wouldn't let me go down the express
elevator, [so] this man agreed to come to my aid. [We] started our
descent on the steps and continued to the 40th floor where he knew of
another elevator bank. The man who was helping me announced to the
people around us that he was helping somebody (me) and that we would
stop at each landing and get out of their way so that they could pass us
easily. We exited the staircase on the 40th floor and found and entered an
elevator. Two women joined us because they thought we knew what we
were doing. We rode the elevator from the 40th floor to the lobby. It
made two stops before expressing to the lobby." Interview 1000048
(NIST 2004)
Some occupants were successful in using an elevator to get to the lobby, but were turned around by
building security, presumably in an effort to minimize the impact WTC 2 occupants would have posed on
the evacuation of WTC 1 occupants:
"I was in an elevator going up to my floor when Building One was hit.
When I got to my floor, we took an elevator back down to lobby, but the
guard sent the elevator full of people back up to the floor." Focus
Group #3 (NIST 2004)
NIST NCSTAR 1-7, WTC Investigation 97
Chapter 6
6.2.4 Announcements in WTC 2
At 8:58 a.m., WTC Channel 15 (NYC EMS Direct Line) recorded a statement from an unknown source,
stating, "I want to start a building evacuation, Building One and Building Two." (PANYNJ 2001a). The
timing was proximate with announcements in WTC 2, made from the lobby fire command station, first
contradicting the Channel 1 5 statement, then supporting it.
At 9:00 a.m., approximately three minutes before the second attack, the first building-wide public address
system announcement was made to the occupants of WTC 2. Synthesizing the content of the
announcement as reported by many interview respondents indicates that the general announcement went
as follows:
There is a fire condition in WTC 1. WTC 2 is secure. Please return to
your offices.
While some occupants interpreted the announcement as less of an instruction and more of a suggestion
that it was safe to return to their offices, others interpreted the announcement as an instruction to remain
where they were or to return to their offices. No respondent reported that the announcement told them to
evacuate.
Approximately two minutes later, at 9:02 a.m., one minute before WTC 2 was attacked, a follow-up
announcement was made, contradicting the previous announcement. Recorded in the background of an
answering machine recording an occupant calling home from a floor in the 90s and subsequently provided
to the interviewer during the interview, the announcement indicated that it was now permissible to initiate
an evacuation:
"May I have your attention please. The situation is in Building 1.
However, if conditions on your floor warrant, you may wish to start an
orderly evacuation." Interview 3000001 (NIST 2004)
The 9:02 a.m. announcement was also noted by an occupant catalogued in the collection of published
accounts (Fahy and Proulx 2003). A survivor in the 70s of WTC 2 was still on his or her floor when the
9:02 a.m. announcement was given over the public address system.
"I stayed on the floor [to] wait until everyone was cleared out. Most of
the people on my floor evacuated using the elevators, which they took to
the skylobby. It seemed to me [that] I would be in more danger by going
out. I heard the phone ringing off the hook and my boss and I were
going from phone to phone answering them to reassure the families that
everyone was ok." Interview 1000524 (NIST 2004).
It was then that United Airlines Flight 175 crashed into WTC 2.
98 NIST NCSTAR 1-7, WTC Investigation
Chapter 7
September 11, 2001, 9:02:59 a.m.
Flight 175 Crashes into WTC 2
"On the phone, [the occupant from a floor in the 90s in WTC 2] kept
describing the people jumping out of WTC 1. I could not hear any noise
in the background; it was quiet with some slight conversations going on.
When the second plane hit WTC 2, I then heard people screaming, "No,
no. Oh my God, no!" Screaming and then silence. People's sounds
disappeared. I could hear 'whopping noises. ' The sounds were similar
to wind blowing and I could hear the fire alarms going off. Then,
nothing else. Just silence. " Interview 300003 (NIST 2004)
, im unni i irtHii i in'" ^^ *
Stairwell A
E
E3E3E3'^U
2=3 = £J=> ■ ■••
.A
n2Tl
Stairwell
B *
At 9:02:59 a.m., a second hijacked Boeing 767-200ER, United Airlines Flight 175, struck WTC 2,
damaging nine floors, from 78 to 84, as shown in Figure 7-1. By the time WTC 2 was hit, 21 percent of
the eventual survivors had already exited WTC 1,
and 4 1 percent of the survivors had already exited
WTC 2. With the second attack, occupants in
WTC 2, like those from WTC 1, began a full-
scale building evacuation.
Alarm tones and public address announcements
were overheard in 9-1-1 audio records from
various floors from floors 82 to 1 05 after the
impact, indicating that some power was available
to most of the building.
As when WTC 1 was struck, a significant surge
in 9-1-1 calls occurred when WTC 2 was struck,
as bystanders and people watching television
called to notify authorities that they witnessed the
event. Many witnesses immediately reported that
a deliberate terrorist attack was underway. Once
again, the WTC 2 attack was described to
operators in a variety of ways, from a large
military aircraft to a bomb, to a commercial
aircraft. 46
Stairwell C
^
©
WTC 2
Figure 7-1. WTC 2 damage (computer
simulated) at 0.62 s after impact with
stairwells superimposed.
New York City 911 Emergency Call Recordings, 2001.
NISTNCSTAR 1-7, WTC Investigation
99
Chapter 7
In WTC 1 , many occupants only had passing knowledge of the second plane strike from cell phone or
electronic communications received by other occupants within the stairwells during their egress:
A survivor from the 50s in WTC 1: "Very soon after entering the
stairwell, someone with a Blackberry communication device received
word that a corporate jet had hit the building. Later the info was clarified
as to what kind of plane it was. Later a report of the second plane hitting.
We could feel the building shake a little, not that strong though."
Interview 1000015 (NIST 2004)
Other occupants sensed that something happened, but had no idea what the source of the disturbance was.
A survivor in the 60s in WTC 1 : "I felt the air clear up due to a big gush
of wind that cleared the smoke (this happened when Building 2 was hit
by the plane)" Interview 1 000 1 00 (NIST 2004)
Occupants of WTC 1 with a view of the plaza observed large pieces of debris, some flaming, crashing
down onto the central plaza. Evacuees from Stairwells A and C used escalators on the east side of the
Mezzanine to proceed down to the lobby. The escalator access point, which often had a group of people
waiting their turn to use the escalators, had a direct view of the debris, as shown in Figure 7-2.
Figure 7-2. East-looking view through WTC 1 lobby as debris from WTC 2 impact travels
past lobby windows onto plaza.
100
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 9:02:59 a.m., Flight 175 Crashes into WTC 2
As occupants exited the building onto the street or plaza level, many were unknowingly walking into
danger, as debris rained down from the impact region in WTC 2. An occupant ofWTC 2 from the 100s,
having used an elevator to get to the lobby, exited WTC 2 at 9:03 a.m.:
"I was guided by what I think were police officers who directed the
crowd. I imagine they were guiding [us] away from where the debris
was falling off Building 1. I walked out straight ahead towards the
Liberty St. exit. Outside of Building 2, debris was falling down when the
second plane hit. A piece of metal went into my arm. I went into shock
due to severe blood loss [and] was assisted by NYPD, which [took] me
to the hospital." Interview 1000563 (NIST 2004)
Due to the angle of the aircraft impact, large regions of survivability remained on some of the floors
directly damaged by Flight 175. Figure 7-3 shows the individual and aggregated damage areas for floors
78 - 83 in WTC 2. See NIST NCSTAR 1-2 and NIST NCSTAR 1-6 for additional building damage
information. From the time of impact and for the next 20 minutes, occupants called 9-1-1 from the
impact region and above to request assistance. 47
Floors 78-83
Cumulative Damage
Severe Floor Damage
Flreprooflng
and partitions ' — '
Floor system i 1
structural damage | |
Floor system
removed
Column Damage
Severed Q
Heavy Damage ^j
Moderate Damage Q
Light Damage
j~7 7 7 7 7 ~ 7 " 7 " "
"T i T T.I
—
l'i i x
n —
"
-
L
A.±.
J—
, — _
4 -I 1 — 1 — —
_
„
J-i-4
_
_
>_. — — -
■
n
.* *•
■
ff 1 '
-
'
-
in
B"t
"j!l
4
'
- i j. *
» *
-
j j-
■r
■
■J
~
> P° 0(
■ <
1
_
m*.
-
"
■ *"
....
Floor 78
Floor 79
Floor 81
Floor 82
■
-
il'i M i!iJ
TTS 1
-
-
Xi ,
-
1 ~ x ~wQ
I
-.','.'.',
■yc=
-
>r i'tt'
, ;:; Ei
■
-•
Floor 80
ma
A H * w -
IUMB
Floor 83
Figure 7-3. Calculated damage to floors 78 through 83 in WTC 2 (see NIST NCSTAR 1-6).
New York City 911 Emergency Call Recordings, 2001.
NIST NCSTAR 1-7, WTC Investigation
101
Chapter 7
Standing on a floor in the 70s in WTC 2, a survivor described the immediate aftermath of the aircraft
impact:
"I heard this deep thud and the ceilings and walls started to crumble. I
grabbed my laptop, ran out of my office. I ran towards Stairwell A
[because] this was the closest stairwell to my office. One of my
colleagues opened the door to Stairwell A, but we didn't go in and turned
around and went south. Stairwell A (the NW stairwell) was pitch black,
filled with heavy black smoke. There was falling debris from the ceiling
making the other two stairwells inaccessible [so we] ran back to
Stairwell A [because] that was only stairwell we had access to. There
was a lot of soot and dust in the air on the way back to Stairwell A. I had
to cover my mouth and nose with my shirt to be able to breathe."
Interview 1000625 (NIST 2004)
A survivor from a floor in the 100s in WTC 2 made it to the 78th floor skylobby just prior to the impact
of flight 175:
I was walking down the stairs, and got off at 78 (stairwell C). I
encountered a lobby full of people (found some people from my group).
The plane hit the building. I went flying and I landed on my right arm.
When that happened my body turned over onto its left side and was
sliding towards the elevators that were being repaired. I thought I would
go down into the elevator shaft (since the door was slightly ajar and you
could see fire coming up and I thought that it would be the end). I
stopped and I called out to my friends and they called back with their
location. I walked over to them climbing over dead bodies. We made an
assessment of what had happened (my boss was dead, other man had
broken legs and my coworker was missing). I got up and was walking in
the direction of the plane. As I was walking there I came across people
who were upset and I was climbing over bodies and I came to realize
there was no communication desk." Interview 1000562 (NIST 2004)
The majority of WTC 2 occupants were not on their usual work floors when WTC 2 was attacked; most
people had initiated or completed evacuation. Therefore, relatively few observations of damage to the
floors in WTC 2 were reported. Figure 7-4 shows a summary of the reported damage to floors in WTC 2.
The information to develop Figure 1-^ was compiled from every source available to the NIST
investigation, including interviews, published accounts, transcripts of emergency communication
channels, and emergency 9-1-1 calls. Note that the "?" denotes a floor where there was no information
found to record the absence or presence of observations. Further, the absence of an observation on any
floor does not positively exclude the presence of that condition as it may simply not have been reported.
1 02 NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 9:02:59 a.m., Flight 175 Crashes into WTC 2
4.
!?
?
■>
■>
?
■>
?
?
3
?
7
->
X
X
X
g
X
g
§
X
g
X
X
X
X
X
S!
X
s
g
X
§
X
X
?
?
?
?
?
?
?
?
?
?
?
?
->
?
?
9
?
?
?
?
?
?
■>
■5
?
->
?
?
?
?
■?
->
o
■p
?
■>
?
?
•>
?
?
->
7
?
?
•>
?
?
?
?
?
f
->
■>
?
•>
?
?
?
?
?
->
->
■>
?
->
?
?
■>
?
7
->
?
7
?
•>
?
?
7
?
7
?
?
?
?
•>
?
?
?
?
■?
?
?
?
?
?
?
?
?
?
?
?
->
■p
?
■>
?
?
■p
?
■p
9
o
?
?
■>
?
?
?
?
?
■?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
? 7 1 7
?
?
/#
'
/v
/■::■
■ : y
Pv
-
=
-
'
7
■>
~r
■>
7
>
■>
?
'
'
»
>
■>
■>
'
7
1
»
'
'
7
■>
>
■>
■>
5
'
'
?
?
'
'
?
'
?
■>
'
'
■>
»
»
'
'
3
£
$
x"
£
g
Si
S
X"
3
g
LJ.
Figure 7-4. Observations of damage from tenant spaces in WTC 2.
NISTNCSTAR 1-7, WTC Investigation
103
Chapter 7
This page intentionally left blank.
104 NIST NCSTAR 1-7, WTC Investigation
Chapter 8
September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m.
56 Critical Minutes
With the attack on World Trade Center (WTC) 2, the events in both WTC 1 and WTC 2 unfolded
similarly, with survivors in both buildings trying to or continuing to evacuate the buildings. For those
trapped above the plane strike in WTC 1, there was no way out. For those above the impact in WTC 2, a
single intact stairwell provided a critical lifeline for at least 18 occupants fortunate enough to both find
and use it. 48
In both buildings, stairwells became not only a means of egress for occupants, but also the only
significant route into the building for emergency responders. Interview analysis found that, prior to the
collapse of WTC 2, the six stairwells in the two towers were relatively passable below the impact regions.
While many occupants indicated that the conditions in the stairwells were better than the conditions
experienced during the 1993 bombing, water, crowding, and the smell of jet fuel were cited as stairwell
obstacles. All the while, building officials and emergency responders tried to cope with an incident of
enormous scope and provide for as orderly an evacuation as possible.
By the time WTC 2 was hit, 91 percent of the survivors in WTC 1 and 87 percent in WTC 2 had begun
their evacuation. Over 20 percent of the survivors in WTC 1 and more than 40 percent in WTC 2 had
already left the buildings at the time of the second airplane strike. Table 8-1 summarizes when survivors
began their evacuation and when they left their respective building.
Table 8-1. Estimated times for survivors leaving the building in WTC 1 and WTC 2.
WTC 1 (n=440)
WTC 2 (n=363)
Begin
Evacuation
Leave Building
Begin
Evacuation
Leave Building
Before WTC 2 hit
91 %
21 %
87%
41 %
After WTC 2 hit, but before
WTC 2 collapse
7%
67%
13%
58%
After WTC 2 collapse
0%
9%
n.a.
n.a.
Key: n.a., not applicable; question not included in telephone survey.
Note: Numbers may not total to 100 percent. Some survivors were unsure when they began their evacuation and/or when
they left the building.
Source: NIST WTC Telephone Survey Data.
See Section 8.1 for further information.
NISTNCSTAR 1-7, WTC Investigation
105
Chapter 8
8.1
SURVIVORS FROM AT OR ABOVE FLOOR 78 IN WTC 2, AFTER IMPACT
Table 8-2. Location of WTC 2 survivors at or
above floors of impact at 9:03 a.m.
Above the floors directly damaged by
Flight 175 in WTC 2, smoke traveled rapidly
upwards. Within twelve minutes of WTC 2
impact, a smoke layer was descending on
floor 105, 49 where multiple callers reported
that between twelve and one hundred people
were trapped. 50 Through face-to-face
interviews and analysis of published accounts,
however, NIST identified 18 individuals at or
above the floors of impact in WTC 2 after
9:02:59 a.m. who escaped. The majority of
these survivors were on the 78th floor, often
referred to as the skylobby, waiting for an
express elevator to the lobby or choosing a
course of action just prior to the impact.
Table 8-2 shows the location of the 18
individuals at 8:46:30 a.m., as well as
9:02:59 a.m. For the hundreds of occupants
who remained above the skylobby when
WTC 2 was hit, two of the buildings three
stairwells had been destroyed by the impact.
Only the staircase furthest away from the
plane's impact was passable (Stairwell A) and
even this stairwell was severely damaged in
places. According to eyewitness accounts,
there was drywall and other debris scattered
over the stairs, water was running down the
stairs, presumably from the sprinkler system
(Murphy 2002), and smoke filled the stairwell
(Adler 2002). In the region near the airplane
impact, the stairwell walls had collapsed, and
occupants had to crawl over or under the
debris in order to descend below the 78th
floor (Adler 2002). According to NIST
interviews and several published accounts, the
conditions in the stairwells then improved significantly below the 78th floor.
Published accounts indicate at least two individuals were below the 78th floor in WTC 2 at the time of
impact, but ascended in order to help injured or trapped colleagues (Fahy and Proulx 2003;
Murphy 2002). Both individuals were building occupants, and one individual was reported to have had
Location at 8:46 a.m.
(WTC 1 Impact)
Location at 9:03 a.m.
(WTC 2 Impact)
1
Floor 103
Floor 78
2
Floor 103
Floor 78
3
Floor 102
Floor 78
4
Floor 101
Floor 78
5
Floor 100
Floor 82 (Stairs)
6
Floor 100
Floor 78 (Elevator)
7
Floor 100
Floor 78
8
Floor 97
Floor 78
9
Floor 97
Floor 78 a
10
Floor 86
Floor 78
11
Floor 86
Floor 78
12
Floor 84
Floor 84
13
Floor 84
Floor 84 b
14
Floor 84
Floor 84 (Elevator)
15
Floor 81
Floor 81
16
Floor 81
Floor 81
17
Floor 79
Floor 78
18
Floor 78
Floor 78
While this individual did evacuate the building, the occupant died
several days later as a result of injuries sustained on September 11,
2001.
After impact, this occupant ascended to floor 91 for a period of time
before finally making it below the impact zone.
New York City 911 Emergency Call Recordings, 2001.
' New York City 911 Emergency Call Recordings, 200 1 .
106
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m., 56 Critical Minutes
an emergency response background. Neither individual successfully evacuated, although several
survivors credited one of the individuals with saving their lives.
8.2 PROGRESS OF EVACUATION BELOW THE IMPACT REGION
For nearly all occupants below the floors of impact, the stairwells became the only means of egress from
both towers. For some the egress went smoothly; for others it was far more difficult. Deteriorating
conditions, coping with crowded stairwells, and the sheer physical effort required to descend dozens of
flights of stairs presented a challenge for many occupants.
A survivor who was on the 78th floor skylobby in WTC 2, then used the
stairs down from there: "People were having general conversations,
seemed calm, and walked at a steady pace, no sense of panic." Interview
1000825 (NIST 2004)
A survivor who began on a floor in the 30s of WTC 1: "On about the
29th or 30th floor, as I was leaning against the wall letting an injured
person pass, I felt the wall heave (which apparently was due to the
second plane hitting the other tower). This was about 9:10. On the stairs,
near the 28th floor, there was a pile of shoes that accumulated from
people kicking them off. Some of the people around me were tripping on
them and warning others to watch out for them. Our speed of descent
was very slow. There were many firemen clogging the stairs and spilling
out onto the 28th floor. They were resting and taking care of one who
was sick—was having a heart attack. They were also carrying much
equipment. All of this was happening at about 9:20 am." Interview
1000042 (NIST 2004)
Occupants also sought information from emergency responders by telephone. Advice from 9-1-1
operators was often to remain in place and await rescue. 5
At 9:20 a.m., also in WTC 1, occupants frustrated by a lack of information reached out to authorities,
including the Port Authority Police Desk.
Male: Oh, hi. We're on the 39th floor here. We're not getting any messages. Can
somebody tell us what's up?
PAPD: Come on down. Just come on down, everybody just come on down. Get down
the staircase. Don't take the elevator.
Male: Is there smoke in the stair?
PAPD: Check the stairway. We're not sure yet.
Male: We've got...
Male: We're in one, 39. We've got, uh...
PAPD: You should be low enough to get down, because it happened up high.
Male: All right.
Male: Thank you. (PANYNJ 200 la)
SI
New York City 911 Emergency Call Recordings, 2001.
NIST NCSTAR 1-7, WTC Investigation 1 07
Chapter 8
As the two communication records exemplify, calls to PAPD from occupants, in general, resulted in
clearer, more specific information, such as where the emergency was. Further, some PAPD operators
quickly recommended to occupants that they should leave the building, including WTC 2 prior to 9:02:59
a.m., although that was not universally true. FDNY dispatch and 9-1-1 operators generally advised
occupants to shelter in place and await rescue, consistent with existing protocols, but not the preferable
course of action on September 11, 2001.
These quotes and communication records point out several issues that occupants had to deal with during
their evacuation. The environment in both buildings continued to worsen with time. In addition to the
physical exertion required to descend the stairways, occupants were also faced with the need to allow
injured persons to pass and firefighters and other emergency responders to travel up the stairs to respond
to the fires. In addition, information from co-workers and managers at times conflicted with local and
building-wide announcements about the event. These issues are discussed in more detail in Chapter 10.
Occupants of both buildings faced a number of challenges in their attempts to evacuate once the planes hit
both buildings. As building authorities and emergency responders worked to organize the evacuation
under rapidly changing conditions, many occupants were directed out of the stairwells during their
evacuation. Thirty-four percent of respondents to the telephone survey reported leaving a stairwell at
least once during their evacuation. While this occurred from floors 3 through 78 in WTC 1, the largest
percentages were found in the skylobbies on floor 78 (7 percent of telephone interview respondents) and
floor 44 (9 percent), as well as on floor 13 (9 percent). Similarly, in WTC 2, the skylobbies on 78
(13 percent of respondents) and 44 (25 percent) were mentioned most often. Some occupants left
stairwells at the skylobbies or other locations to seek additional information or to find a better route out of
the building. A survivor from a floor in the 90s in WTC 1 used multiple stairs during egress:
"The staircase became crowded at about the 68th floor and I hoped to
find a less crowded stairwell. I descended the stairs from 78 to about the
50s, before exiting the stairs again. The stairs at 68 became more
crowded because they were evacuating some injured or handicapped
people. The speed slowed to a halt— so we exited to find another
staircase. I entered staircase B (in the 50s), and descended to the 40s
before exiting the stairs again to find better conditions as I exited the
building." Interview 1000052 (NIST 2004)
Occupants were also directed to leave the stairwells.
A survivor who began on a floor in the 80s of WTC 1 : "At the skylobby,
we were re-directed by Port Authority personnel. We went to the
stairwell on the far west side of the building. I saw many people trying to
get into the little door of the stairwell. I went back to a different stairwell
that was empty. The center stairwell was the one we ended up in that
took us to the ground floor lobby." Interview 1000535 (NIST 2004)
In WTC 1, occupants with mobility impairments presented a particular challenge. A specific floor was
designated by emergency responders to hold mobility-impaired occupants. The idea was to allow fully
ambulatory occupants to egress, and then move the mobility-impaired occupants out once the stairwells
had cleared somewhat. Numerous interview respondents indicated that a relatively small number of slow-
moving occupants hindered the flow of faster moving occupants behind them in the stairwells. Reports of
108 NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m., 56 Critical Minutes
the specific floor number in WTC 1 varied (12, 18, or 20), but several survivors recalled the existence of a
holding floor.
A wheelchair-user from a floor in the 60s was being assisted by four
previously unknown occupants down the stairwells in WTC 1 : "We saw
a Port Authority person. He told us to go to the 18th floor [because]
there was an evacuation station for people who needed special assistance.
I was carried mostly by the two [occupants] who wandered onto my floor
off the stairs [in the 60s], but then there were two more that also helped.
They switched off. At about the 30th floor, we encountered firemen for
the first time. We moved to the right." Focus Group #3 (NIST 2004)
NIST found no evidence that a similar holding floor for mobility-impaired occupants existed in WTC 2.
8.3
EVACUATION ROUTE THROUGH CONCOURSE
The first firefighter killed at the WTC was reportedly killed by a falling person (FDNY CD 12/15 2001);
many evacuees and bystanders were injured by debris. Figure 8-1 shows the WTC plaza from the
WTC 1 Mezzanine Level, perhaps viewed by thousands of occupants as they prepared to walk down the
escalators to the Concourse Level. Throughout the evacuation, the lobbies and particularly the Concourse
area served as the primary
route to safety for survivors
of the WTC attacks. Within
approximately 1 minutes of
the attack on WTC 1,
building officials and police
in WTC 1 and WTC 2 had
established a procedure to
route evacuees from both
buildings through the
Concourse (mall) rather than
the plaza in order to prevent
occupants from being
injured by falling debris or
people or being upset by
shocking sights. As
presented in Section 2.2.2,
Stairwells A and C in both
WTC 1 and WTC 2
discharged to the Mezzanine
Level, while Stairwell B
discharged to the Concourse Level, without the choice of exiting at the Mezzanine Level. Occupants who
used Stairwell A or Stairwell C were directed to use an escalator (see Figure 8-2) in order to descend to
Figure 8-1. View from mezzanine level in WTC 1 looking east
across WTC plaza, covered with debris. 'Sphere,' artwork by
Fritz Koening, can be seen on the plaza.
NISTNCSTAR 1-7, WTC Investigation
109
Chapter 8
the Concourse Level. Many face-to-face
interviewees who used Stairwell A or C
reported that the escalators were turned off
in order to allow occupants to descend
easily using both escalators. 52
Once through the underground shopping
mall, an occupant originating from a floor
in the 30s in WTC 2, recalled being
directed out of the Concourse at
approximately 9:03: "The security guards
were like a human chain telling us which
direction to go. We followed the security
guards direction, half walking, half running
towards Borders [retail book store] to now
go up from the Concourse to get out."
Interview 1000842 (NIST 2004) Ascending
the escalator from the shopping complex
was captured in Figure 8-3.
Source: John Labriola 2001, edited by NIST.
Figure 8-2. Escalator from mezzanine to
concourse level in WTC 1 on September 11, 2001.
8.4 MORE ANNOUNCEMENTS
IN WTC 2
At 9:12 a.m., the fire command station in
WTC 2 (310-B) used WTC Ch. 22 to
broadcast that the fire (floor) warden
phones in WTC 2 were not working. 53
Further, WTC personnel indicated that
they were making announcements that
wardens should not wait for further
instructions over the floor warden phones,
as per protocol.
Figure 8-3. Evacuees leaving complex near WTC 5.
310-B to any units: Be advised that Building Two (inaudible) only
(inaudible) warden phones. We can't pick up warden phones. We are just
making straight announcements telling the people not to stay at the
warden phones, because we can't pick them up. (PANYNJ 2001a)
NIST found no other evidence to confirm or refute that the fire command station issued such an
announcement or that any occupant heard an announcement around that time.
Analysis of Face-to-face Interviews (NIST 2004).
For an explanation of the emergency communication channel assignments, see NIST NCSTAR 1-8.
110
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m., 56 Critical Minutes
At 9:20:10 a.m., approximately 17 minutes after the second tower was hit, the fire command station made
another announcement over the public address system updating the occupants. The announcement was
recorded in the background of multiple 9-1-1 calls. 54 To the best of NIST's ability to understand, the
announcement went as follows:
"This is the fire command station in building number two. We are in the
lobby. We are holding down here. The condition seems to have
subsided somewhat. There are people in the building. There are some
people have left. If you wish to leave, you can now use the Concourse.
You may walk to the Concourse on this side, avoiding building number
one. The condition seems to have subsided in the lobby of building two.
Some people have left the building. We are here monitoring the
situation. The fire department are concentrating their efforts on building
number one. People are leaving the building...
(unintelligible). . .Number 2 . . . (unintelligible) ..."
The announcement lasted between 60 and 70 seconds, with the unintelligible portion comprising
approximately the final 10 seconds. The content of the 9:20 a.m. WTC 2 public address system
announcement varied significantly from the pre-planned building evacuation announcement covered in
Section 2.2.4. However, no survivors from whomNIST collected first-person accounts reported hearing
the 9:20 a.m. announcement in WTC 2. This may be attributable to the fact that approximately 75 percent
of all WTC 2 survivors had left WTC 2 by 9:20. Two occupants, trapped above the floors of impact,
however, did respond to the 9:20 a.m. announcement, according to 9-1-1 records. 55
8.5 CONDITIONS CONTINUE TO DETERIORATE IN BOTH TOWERS
As the events unfolded, conditions within both towers continued to worsen. At or above the floors where
the airplanes hit the buildings, occupants had to cope with fire, smoke, and severe building damage. In the
time period after 9:15 a.m., only four calls to 9-1-1 came directly from inside WTC 1. As Figure 8-4
shows however, many occupants above the impact region in WTC 1 remained alive at 9:58:59 a.m.,
shortly before WTC 2 collapsed.
Below the impact floors, occupants worked to evacuate the building while emergency responders made
efforts to organize the evacuation and gain access to damaged floors to rescue injured or trapped
occupants. At 9:35 a.m., a WTC Official in the security command center (SCC) reported to S2 trouble
unlocking the remotely-controlled doors:
MaleB:S2toSCC.
Female A: Go, S2.
Male B : How are you doing up there?
Female A: We've got a lot of smoke.
Male B: We're on 16 now.
Female A: That's a copy. We can't use software right now to try to release the doors.
(PANYNJ2001a)
54 New York City 9 1 1 Emergency Call Recordings, 200 1 .
New York City 911 Emergency Call Recordings, 200 1 .
NIST NCSTAR 1-7, WTC Investigation 1 1 1
Chapter 8
Figure 8-4. Occupants seeking fresh air on floors 103 - 105 on north face of
WTC1 at 9:58:12 a.m.
The reported difficulty unlocking doors did not appear to directly prevent any occupants from
successfully evacuating, however, as every fourth door was always open and the doors were only locked
112
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m., 56 Critical Minutes
from the stairwell side. Further, only the mechanical rooms and roof were controlled electronically. The
communication does reflect the condition of the SCC and the difficulty the conditions posed for
operations originating from that office. Numerous other reports of fire, smoke, and damage were noted in
transcripts of PANYNJ radio calls regarding the 22nd floor command center (PANYNJ 2001a).
In WTC 1 , more than half of the survivors noticed smoke in their immediate area during their evacuation
compared to only 10 percent who reported noticing smoke at first awareness. At least 25 percent reported
collapsed walls during their evacuation compared to only 6 percent who noticed collapsed walls at first
awareness.
Table 8-3 shows a summary of observations by the survivors during their evacuation from WTC 1 and
WTC 2. All percentages increased from earlier observations on their floors prior to beginning evacuation.
Table 8-3. Observations of conditions during evacuation of WTC 1 and WTC 2.
Observation
WTC 1 (n=440)
WTC 2 (n=363)
Smoke
57%
21 %
Sprinklers/water
55%
3%
Fatally injured people
41 %
8%
Power outage/flickering lights
37%
29%
Jet fuel
31 %
21 %
Fallen ceiling tiles
29%
12%
Fire alarms
26%
19%
Collapsed walls
25%
11 %
Extreme heat
10%
7%
Fire
9%
4%
Fireballs
3%
2%
Conditions in the stairwell were challenging, as well.
A survivor from a floor in the 20s in WTC 1 : "The stairwell was lit the
entire way down. There was a grayish color smoke which smelled like
fuel. The more we reached the lower floors the stronger the smell
became. On the 6th floor, the sprinklers were on, which slowed us down
because we wanted to be cautious and not slip or fall." Interview
1000044 (NIST 2004)
An occupant who survived the horror of aircraft impact on the 78th floor
in WTC 2:"... not much debris, but there was a pipe that we had to go
under, hanging wire, and water in the stairs; no counter flow after we
were hit, stairs were well lit, but 78 had no lighting. On 40-well lit, a lot
of doors were locked, stranger smell that I can't identify. As we were
walking down, we saw fireman coming up and told us to get to 40 and
that someone would take us in an elevator down to the lobby." Interview
1000562 (NIST 2004) Firefighters and security personnel would escort
NISTNCSTAR 1-7, WTC Investigation
113
Chapter 8
the occupant down in the elevator, out through the Concourse area to a
waiting ambulance.
At 9:36 a.m., an occupant's telephone call to a New York City 9-1-1 telephone operator indicated that a
floor in the 90s of WTC 2 had collapsed. According to NYPD records, information from this call
concerning the floor collapse appears to have been conveyed inaccurately by the 9-1-1 call taker and the
NYPD radio dispatcher. 56 The NYPD dispatcher transmitted the message at 9:41 a.m. and 9:51 a.m.
Independent of the exact floor number, the call content demonstrated deteriorating conditions above the
impact region in WTC 2.
Table 8-4 shows the total evacuation time for survivors in WTC 2. Total evacuation time was measured
from first awareness to exiting the building and included time spent on the starting floor. While the total
evacuation time increased from the lower floors to the middle floors, as would be expected when the
proportion of occupants who used elevators was roughly constant, the total evacuation time for survivors
from the upper floors decreased due to the increased proportion (approximately twice as many, see
Section 6.2.3) of survivors who used elevators to evacuate.
Table 8-4. Total evacuation time (min) for survivors from WTC 2.
Total Evacuation Time" (min)
25%
Evacuated
50%
Evacuated
75%
Evacuated
Average Time
(min)
Lower floors (Basement - 42)
10
15
21
17.6
Middle floors (43-76)
23
35
45
34.9
Upper floors (77-110)
11
23
36
25.5
All survivors in WTC 2
13
21
35
25.0
a. Total evacuation time was measured from first awareness to exiting the building.
Note: Italics indicate the presence of a bias in the underlying data. Interpretation and comparison of data should consider
this bias.
Source: NIST WTC telephone survey data.
Minutes prior to the collapse of WTC 2, an NYPD Emergency Services Unit (ESU) officer radioed from a
floor in the 20s to the outside that he was having trouble ascending the stairwell due to the large number
of occupants descending (Interview 24 NYPD [NIST 2004]). While the origin of the occupants remains
unknown, only 1 1 occupants who started evacuating below the impact region were known not to have
survived.
WTC 2 collapsed at 9:58:59 a.m., leaving over 600 occupants and over 100 emergency responders dead.
NYPD, McKinsey and Company, NYPD call-routing and message dispatch, "106' floor is collapsing." Draft summary report
July 23, 2002.
114
NIST NCSTAR 1-7, WTC Investigation
Chapter 9
September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m.
Collapse
"The lights were off and it was dark. You could barely see your hand in
front of your face and there was a lot of dust. We just stood there and
waited for the dust to settle. The emergency lights went on and we tried
resuming going back down those stairs for half a flight. People [were
coming back] up the stairs and said, 'the egress is gone.'" Interview
1000594 (NIST 2004)
At 9:58:59 a.m., many of the 624 to 680 occupants and building personnel who had fought for survival in
WTC 2 for the 56 minutes after the building was attacked were killed by the building collapse. Over 100
emergency responders, unaware of impending collapse, were killed as well (NIST NCSTAR 1-8 2005).
The effect of the collapse of WTC 2 was felt not only at the remaining WTC buildings, but all over Lower
Manhattan, as shown in Figure 9-1.
9.1 EVACUATING WTC 1
Although occupants in WTC 1 may or may not have comprehended the magnitude of the collapse at the
time, it still had a significant impact on the evacuation of remaining survivors in WTC 1 . As described
below, survivors related to NIST, during face-to-face interviews, the nature of the cloud of debris and
smoke which filled the Concourse and lower portions of the stairwells in WTC 1 , making evacuation that
much more difficult for those still left in the building.
A survivor in the shopping Concourse who began on a floor in the 50s of
WTC 1 : "We heard (what we later learned) the South Tower/WTC 2
collapsing. We stopped near the joining of the Gap store and the PATH
escalators. The sound, at first, was muted and rumbling. It was a
scraping sound that grew louder over the course of 15-30 seconds. It
seemed like it was going to envelop us. The sound became very loud
(like steel on steel). Very quickly after WTC 2 collapsed, the lights went
out in the Concourse. Immediately, people screamed and panicked, and
started to run in the dark. This was the one time I thought I really was
going to die. I saw what I thought then was a wall of water, like a
tsunami, coming at me (I thought I was going to drown). What it turned
out to be was the cloud of debris and smoke. It irritated my eyes and
filled my mouth, and made me cough and spit. I called to the others (my
staff/ co-workers) who were with me. I felt the need to be in touch with
them, since we had come all this way together." Interview 1000106
(NIST 2004)
NIST NCSTAR 1-7, WTC Investigation 1 15
Chapter 9
©2001. N ew York C i ly Police De pa r t m ent. All rights reserved.
Figure 9-1. Collapse of WTC 2, viewed from north.
A survivor still in a stairwell who began on a floor in the 30s of WTC 1:
"We descended on this stair until we got to the landing between floors 7
and 8. The lights now went out in the stairwell. I entered the 8th floor
and went down a corridor to a different staircase (I'm not sure which
stair, but I believe it was the original one I started out on). People on the
stairs below me (who apparently got a big rush of air) said the floors
below us were collapsing. They turned around and started running back
up and they were covered with a gray, powdery substance, such as ash or
smoke — so we believed them. Then when the fireman opened the door
on the 8th floor and said to come onto the 8th floor, I readily followed
116
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m., Collapse
his instructions. Some didn't follow those instructions and continued
back upstairs. I descended the stairs from the 8th floor to what I believe
was the 5th floor. It was still dark in the stairwell. I changed stairwells
for the last time and descended to the Mezzanine Level. I exited the stairs
onto the Mezzanine, headed toward the west wall, then the north wall,
and finally exited through the emergency exit door." Interview 1000042
(NIST2004)
A survivor from a floor in the 50s in WTC 1 : "We thought it was a bomb
at this point, however, it was Building Two collapsing. Vibrations
stopped; the noise was incredible; the lights went out and the sprinklers
came on. We walked a lot more quickly. We were getting out of the
building. It was hard to walk down the stairs because of the water
coming from above from the sprinkler system. There was a large amount
of smoke and I couldn't see the person in front of me. I guess we smelled
jet fuel. I could barely see the emergency strips on the stairs, but you
could see them [because] the lights had gone out." Interview 1000054
(NIST 2004)
A survivor in the stairwell who began in the teens of WTC 1 : "We heard
a gigantic explosion. We reached the second floor landing and we
stopped. It was pitch black; black and tremendous suffocating smoke and
dust. We stopped there and choked and froze. The lights went back on
after it seemed like forever. We were told by the firemen and rescue
workers that we had to go up to the fourth floor, that the exit was
blocked." Interview 1000093 (NIST 2004)
Whereas occupants were primarily using the cover of the underground Concourse Level to escape the
WTC site prior to the collapse of WTC 2, the debris created by the collapse forced the remaining
occupants to proceed outdoors after they left the stairwells. For occupants leaving Stairwells A and C on
the Mezzanine Level, the exit on the North wall, West corner provided access to the Tobin Plaza, from
which many proceeded North between WTC 5 and WTC 6; others used the bridge to cross West St to the
World Financial Center. Occupants leaving Stairwell B on the Concourse Level moved West to the West
Street VIP entrance, where many continued to the World Financial Center, while others traveled North
along West St.
The evacuation rate slowed dramatically in WTC 1 after WTC 2 collapsed. Clearly, the obstacles
presented by the dust, debris, and loss of power from the WTC 2 collapse contributed to slowing the
occupants travel speed. By 9:58:59 a.m., 88 percent of survivors (over 6,500 people) in WTC 1 had
exited the building. Approximately 900 additional occupants and many emergency responders would exit
WTC 1 over the remaining 29 minutes before the building collapsed.
Many mobility-impaired occupants remained in a holding pattern on a lower floor in WTC 1 during this
time. An FDNY Battalion Chief wrote "Last Man Down," in which he described walking onto the 12th
floor of WTC 1:
"The office was filled with people. Forty, fifty, sixty of them. I didn't
stop to count, but there they were, in all shapes and sizes, of all ages. All
of them were sitting quietly, patiently, apparently waiting for instruction
NIST NCSTAR 1-7, WTC Investigation 1 17
Chapter 9
or assistance... There had to have been a couple dozen nonambulatory
office workers in that room, easy, and each one of them had one or two
or sometimes three friends helping them. Friends, coworkers, people
they'd just met in the stairs, for all I knew." (Picciotto 2002)
Some occupants, however, had resisted the idea of delaying their evacuation by stopping on the
designated holding floor. A mobility-impaired occupant who started in the 60s had been instructed while
in the stairwell by a Port Authority employee to report to the 1 8th floor, when the time came to choose
between resting or continuing to evacuate:
"At about the 22nd floor, Building Two collapsed. The building swayed.
There was smoke and dust from the collapse of building two. I had my
mask on, but it was hard to breathe and see. We picked up speed.
Firemen started coming back down (from above us). We had to leave the
staircase because we couldn't see or breathe. We went into a tenant's
space for 5 or 10 minutes. Then, we couldn't find another staircase so
we went back to the one we had just left. When we got to the 1 8th floor,
we just kept going. The [helpers] asked me if I wanted to stay, I said
"No." I had been with them this far. After that we were moving more
quickly." [NIST Focus Group #3, 2004]
Battalion Chief Picciotto described ordering the evacuation of the holding floor in "Last Man Down,"
largely steering them down Stairwell B, at approximately 10:21 a.m. The fate of those mobility-impaired
occupants remains unclear. The conditions inside the stairwells, after WTC 2 collapsed, became more
difficult.
A survivor who began on a floor in the 60s of WTC 1 : "The staircase
was a mess; it was filled with concrete, pieces of metal sticking out of
the stairs. There were also holes on the walls. There was also lots of
water on the floor." Interview 1000543 (NIST 2004)
While building personnel attempted to provide building-wide announcements in WTC 1 , only two
survivors remembered hearing announcements in WTC 1, both near the time of the collapse of WTC 2.
Specific details of the announcements were not recalled beyond "some sort of an emergency" or "to
evacuate the building" Interviews 1000594 and 1000068 (NIST 2004), respectively.
After WTC 2 collapsed, there was a sharp decrease in 9- 1 - 1 call volume from inside WTC 1 . Above the
floors of impact in WTC 1, NIST found only one call to 9-1-1 from a person inside WTC 1.
Table 9-1 shows the total evacuation time for survivors in WTC 1 overall, as well as by zone (low,
middle, and high). Comparisons of upper zone data with data from lower and middle zones should be
carefully considered, however, as the upper zone represents a smaller region (fewer number of floors and
less vertical height represented) than the lower and middle zones. This was due to the fact that no
individuals survived from above the 9 1 st floor.
Several calls were received from persons relaying information on behalf of occupants in WTC 1, but the time delay in relaying
the information from the occupant to the 9-1-1 dispatcher was unknown.
118 NIST NCSTAR 1 - 7, WTC Investigation
September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m., Collapse
The quartile and mean times for survivors to exit the building shown in Table 9-1 increased as the
survivors were higher in the building. The first quartile of surviving evacuees from the middle zone were
exiting WTC 1 at approximately the same time as the last quartile of surviving evacuees from the lower
zone (approximately 40 minutes). Similarly, the first quartile of surviving evacuees from the upper zone
were exiting WTC 1 at approximately the same time as the last quartile of surviving evacuees from the
middle zone (approximately 60 minutes). Note, however, that the number of floors (and vertical range) of
the upper zone was significantly less that the lower and middle zones. Therefore, comparisons of the
upper zone times to the lower and middle zone times should factor the zone differential into the analysis.
Table 9-1. Total evacuation time (min) for survivors 1
from WTC 1.
Total Evacuation Time"
(min)
25%
Evacuated
50%
Evacuated
75%
Evacuated
Average Time
(min)
Lower floors (Basement - 42)
16
27
42
29.0
Middle floors (43-76)
41
51
65
54.7
Upper floors (77-91)
60
71
77
70.3
All survivors in WTC 1
25
41
58
41.9
a. Total evacuation time was measured from first awareness to exiting the building.
Note: Italics indicate the presence of a bias in the underlying data. Interpretation and comparison of data should consider
this bias.
Source: NIST WTC Telephone Survey Data.
9.2
LOCATION OF VICTIMS AND SURVIVORS WITHIN THE BUILDING
Once the two towers collapsed, the fate of 2,749 building occupants and emergency responders was
sealed. Table 9-2 shows the likely location at time of impact of all decedents identified by the City of
New York as decedents. It also categorizes types of emergency responders who perished in the disaster.
To identify locations, NIST relied on more than 300 face-to-face interviews and 800 telephone surveys,
various Web sites maintained by survivors or victims' families and colleagues, several media outlets'
reports, and a badge list maintained by The Port Authority of New York and New Jersey. Table 9-2
below shows the results of that analysis.
The following sources were used to identify the likely location of decedents at time of impact:
• Septemberl lVictims.com: This site is dedicated to the victims of September 11, 2001 tragedy.
• Portraits: 9/1 1/01 : Published by the New York Times in 2003, this book includes short interviews
with family members of many decedents.
• CNN.com In-Depth Special (http://www.cnn.com/SPECIALS/2001/memorial/index.html):
Tribute site for people to write remembrances of decedents.
• Badge List maintained by Port Authority of New York and New Jersey: Includes name,
employer, building, and floor for all occupants with badge-access to WTC 1 or WTC 2.
• Numerous memorial sites maintained by companies which lost employees: Includes names and
remembrances of decedents. Examples include the Port Authority, Fire and Police Departments,
Marsh & McLennan Companies, EuroBrokers, Fiduciary Trust, and others.
NISTNCSTAR 1-7, WTC Investigation
119
Chapter 9
• Newsday.com: Includes short stories written about specific decedents.
Table 9-2. World Trade Center decedent location.
Likely Location at Time of Impact 3
World Trade Center 1 Occupants
1,462
At or Above Impact
1,355
Below Impact
107
World Trade Center 2 Occupants
630
At or Above Impact
619
Below Impact
11
Confirmed Below Impact in WTC 1 or WTC 2
30
Unknown Location Inside WTC 1 or WTC 2
24
Emergency responders (Total)
421
FDNY
343
NYPD
23
PAPD
37
Hospital/Paramedic
7
Federal
2
Volunteer Responders
9
Bystander / Nearby Building Occupant
18
American Flight 11
87
United Flight 175
60
No Information
17
Total
2,749
a. Where possible, NIST has used eyewitness accounts to place individuals. Where no specific
accounts existed, NIST used employer and floor information to place individuals.
The analysis shown in Table 9-2 generally assumes that individuals were on their work floor when his or
her building was hit. The employer and floor number from the badge list were generally used to
designate location. An employee of Cantor Fitzgerald, for example, would be assumed to have died
above the floors of impact in WTC 1, unless evidence to the contrary was discovered. Often, information
about the activities or location of an individual was available from on of the sources described above, and
that information was used to categorize individuals. Additional information was necessary to place the 18
individuals documented as bystanders or occupants of nearby buildings. There were 30 individuals who
were known to have been below the floors of impact, but whose locations could not be determined.
Typically, these individuals were security guards and fire safety staff who were observed performing
activities below the floors of impact after the airplanes struck. There were 24 individuals who were likely
in the building, but for whom no clarifying information could be uncovered indicating whether they were
above or below floors of impact. Their occupations largely included maintenance, janitor, delivery,
safety, or security functions.
120
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m., Collapse
Emergency responders were defined to be people who arrived at the site from another location; thus,
security staff and Port Authority staff (different from PA Police Officers) were not defined as emergency
responders.
No information could be determined for 17 individuals.
Estimated Number of Decedents below the Impact Regions
The total number of people who perished below the floors of impact was:
421 emergency responders;
107 individuals in WTC 1;
1 1 individuals in WTC 2;
30 individuals below the impact region in either WTC 1 or WTC 2;
Up to 24 individuals were somewhere in the building, but may have been above or below impact
regions in either building;
Up to 17 individuals for whom no information could be located;
Thus, there were between 569 - 610 individuals who were below the impact floors but did not
successfully evacuate. Assuming that all of these individuals were alive at the time that his or her
respective building collapsed, approximately 20 percent of those lost on September 11, may have been
alive just prior to collapse of the WTC towers.
Estimated Number of Decedents at or above the Impact Region
The total number of individuals who perished who were at or above the floors of impact was:
• 1,355 individuals above the 91st floor in WTC 1 at 8:46:30 a.m.;
• 619 individuals above the 77th floor in WTC 2 at 9:02:59 a.m.;
• Up to 24 individuals were somewhere in the building, but may have been above or below impact
regions in either building;
• Up to 17 individuals for whom no information could be located.
Therefore, between 1,974 and 2,005 individuals were at or above the area of impact in WTC 1 and
WTC 2 who did not successfully evacuate.
NIST NCSTAR 1-7, WTC Investigation 121
Chapter 9
Number of Aircraft Passengers and Crew
The total number of airplane passengers and crew was 147, not including the 10 hijackers. 58 American
Flight 1 1 had 87 passengers and crew, while United Flight 175 had 60 passengers and crew.
Other Victims
Additionally there were 18 people (non-WTC occupants) killed outside the building by debris and jet fuel.
9.2.1
Fatalities in Elevators
USA Today (Cauchon 2001) reported that as many as 200 occupants died in elevators in WTC 1 and
WTC 2. While the NIST Investigation did not produce an estimate of the number of occupants who died
in elevators, NIST did collect information related to elevators and elevator use that may be of benefit in
identifying ways of reducing building occupant risk in general.
The elevators were the normal means of vertical travel in WTC 1 and WTC 2, and it was likely that many,
if not most, of the elevators were in use at the time of aircraft impact. There were two ways in which
lives of building occupants may have been lost on September 11, 2001 while using the elevators. First, if
the cables above the elevator car were severed, the car would have dropped to the bottom of the shaft
(which may not be the basement, depending upon the specific elevator shaft). The safety brakes on the
elevator car would have been activated by the governor cable when the governor wheel exceeded the
design speed. Since the governor was mounted at the top of the hoistway, should the governor cable have
been cut along with the hoist cables, the safety brake would not have been activated, and the car would
have fallen. If the car dropped far enough, occupants of the car may not have survived the impact. For
shorter falls, the occupants may have survived but remained trapped in the elevator car at the bottom of
the shaft.
Each of the towers had 99 elevators serving the above ground floors, but most were locals that ran to only
a few floors within one of the three elevator zones. For an elevator's cables to be cut and result in
dropping the car to the bottom of the shaft, the cables would need to have been in the aircraft impact
debris path, floors 93 through 98 in WTC 1 or floors 78 through 83 in WTC 2. Inspection of the elevator
riser diagram and architectural floor plans for WTC 1 shows that the following elevators met these
criteria: cars 81 through 86 (Bank B) and 87 through 92 (Bank C), local cars in Zone III; car 50, the
freight elevator, and car 6, the Zone III shuttle. Inspection of the elevator riser diagram and architectural
floor plans for WTC 2 shows that the following elevators met these criteria: cars 75 through 80 (Bank A),
local cars in Zone III; car 50, the freight elevator, and car 6, the Zone III shuttle. Bank A local cars could
have fallen a maximum of 6 floors, Bank B cars 16 floors, and Bank C cars 23 floors. Cars 6 and 50
could have fallen all the way to the pit in the sub-basement level, and car 50 in WTC 1 was reported to
have done so. 59 Elevator 81 (Bank B) in WTC 1 was apparently not destroyed, as WTC Radio Channel
25 (maintenance and electric) recorded a request for assistance in getting the elevator doors opened and
noted smoke and water infiltration into the elevator cab (PANYNJ 2004).
New York City has not issued death certificates for any hijackers.
' NIST Interview, 2004.
122
NIST NCSTAR 1-7, WTC Investigation
September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m., Collapse
In WTC 1, several occupants seeking to use elevators reported that the elevators were not working
immediately after impact, despite signage and a training program specifically prohibiting elevator usage
in the event of an emergency (see Chapter 4). In WTC 2, analysis of telephone interviews determined
that approximately 1 8 percent of the survivors evacuated primarily using an elevator on
September 11, 2001. As with WTC 1, however, it remains unknown how many occupants were using the
elevators at the time of impact. In addition to the occupants who may have been killed while using
elevators, several occupants reported numerous fatalities in the 78th floor skylobby, where occupants
were either waiting for elevators or deciding whether to return to their offices when the second aircraft
struck WTC 2. Interview 1000562 (NIST 2004) and (NIST NCSTAR 1-7A).
Not all occupants in an elevator when their building was hit were killed. Some occupants were able to
force doors open, a few 'chopped' their way through walls (one group famously used a window washing
squeegee [Dimarco 2004]), while others were able to climb out of elevators which had fallen to the
bottom of their respective shafts (Cauchon 2001). Figure 9-2 shows occupants leaving an elevator in the
lobby of WTC 1 shortly prior to the collapse of WTC 2 on September 11, 2001. The reason the
occupants were able to escape after an extended entrapment remains unclear, although Port Authority
officials speculate that the elevator may have reverted to slow-speed (50 fpm [0.25 m/s]) run, designed to
prevent entrapment. 60
© 2001 Goldfish Pictures
Figure 9-2. Occupants leaving a previously stuck elevator on September 11, 2001, in
the lobby of WTC 1.
Second, a number of safety systems and sensors in the WTC 1 and WTC 2 elevator systems were
arranged to stop the elevator in place if they detected an abnormal condition (PA Interview 5
[NIST 2004]). Stopping an elevator car generally represents the safest course of action. The intercom in
'Public Comments to Draft NCSTAR 1-7. PANYNJ. 2005.
NIST NCSTAR 1-7, WTC Investigation
123
Chapter 9
every car would then be used to notify authorities of an entrapment. Given the size of WTC 1 and WTC 2
and the number of elevators, however, occupant extraction could have taken hours, as evidenced by the
five or more hour waits for some occupants during the 1993 bombing (Isner and Klem 2003).
In fires and earthquakes, however, entrapment may result in additional risks to the occupants before they
can be removed. In an earthquake, for example, one of the greatest safety risks comes from the
possibility that the counterweight jumps its track and is struck by the car. Seismic switches on elevators
in seismic zones are designed to detect the earthquake and cause the system to stop, move the car in the
direction away from the counterweight to the next floor, stop and open the landing doors (ASME 17.1,
Safety Code for Elevators and Escalators).
Another issue regarding entrapment was the presence of "door restrictors." A door restrictor is a safety
device required on all new elevators (ASME 17.1, Safety Code for Elevators and Escalators) and retrofit
on existing elevators (ASME 17.3, Safety Code for Existing Elevators) upon any elevator equipment
upgrade. Door restrictors mechanically prevent the elevator car doors from being forced open unless the
car floor is within 36 in. of a landing floor. They were installed in some, but not all, elevators in WTC 1
and WTC 2 (Interview PA 5 2005). Door restrictors address the significant safety hazard that occurs
when occupants force the car doors open and fall through the gap down the shaft or get crushed between
the car and shaft wall. Door restrictors have all but eliminated these accidents, but have made self-escape
by entrapped occupants nearly impossible. Elimination of door restrictors alone would lead to increased
deaths from accidents. A better solution would address both issues.
Occupants could have become entrapped at any level if the aircraft impact caused the elevators to stop. If
the car floor were not within 36 in. (plus or minus 18 in.) of a landing floor, the door restrictor would
have prevented the car doors from being forced open from the inside, and the roof hatch could only be
opened from the top of the car, making for self-escape extremely difficult. If the car was within 36 in. of
a landing, occupants could have forced the car doors open and then would have had to force open the
landing doors or break through the elevator shaft wall, which was 3 in. to 3.5 in. of gypsum material.
124 NIST NCSTAR 1-7, WTC Investigation
Chapter 10
Discussion and Analysis
The attacks on the World Trade Center (WTC) towers were unprecedented events, extending well beyond
the typical engineering design scenarios of buildings on many levels. Occupants were forced to deal with
a full evacuation of severely damaged buildings, although training and procedures called only for partial
evacuation near affected floors. Emergency responders and building management were forced to deal
with multiple extreme disasters in the buildings while facilitating the safe egress of some occupants and
attempting the rescue of others.
This chapter provides a discussion of the factors that affected the overall safety of occupants on
September 11, 2001. It is divided into several topics areas: a summary of total evacuation time for
WTC 1 and WTC 2 from the telephone survey data, an analysis of the important factors that affected total
evacuation time and its components of delay prior to evacuation, and evacuation time in stairwells or
elevators. The face-to-face interview data also highlight additional factors that were not included in the
telephone survey, but were prevalent in sufficient number in the face-to-face data to warrant discussion.
10.1 OVERVIEW OF TOTAL EVACUATION TIME
Figure 10-1 shows the percentage of survivors who reported evacuating over five minute intervals,
measured from the moment that WTC 1 was attacked. There was a large peak in evacuation in WTC 2
during the time period immediately surrounding the WTC 2 impact. Recall that WTC 2 was attacked at
9:02:59 a.m., approximately 16 minutes after WTC 1 was attacked. Nearly one in six WTC 2 survivors
reported having left the building in a single five minute period between 9:01 a.m. and 9:06 a.m. Over
40 percent of the surviving occupants left WTC 2 prior to 9:02:59 a.m. That so many evacuated was
largely due to occupant use of the elevators in WTC 2 to evacuate prior to impact. Approximately one in
six surviving occupants (18 percent) initially used an elevator to evacuate WTC 2. After WTC 2, was hit,
only one elevator was operating, under the control of the Fire Department of New York.
Integrating the rates shown in Figure 10-1 yields Figure 10-2, which represents the fraction of survivors
remaining in the building at different points in time. Within the bounds of statistical uncertainty, the
same number of occupants successfully evacuated WTC 1 as WTC 2. Therefore, comparing the slopes of
the curves is appropriate. For a period of approximately 30 minutes in WTC 2 (when only stairwells were
available) the slope (or evacuation rate) roughly matches the slope of the first 70+ minutes of evacuation
in WTC 1 . This result would be expected given the similarity of the egress systems in the two buildings.
The evacuation rate declines significantly for the final twenty minutes before each building collapses, due
to the fact that over 90 percent of the occupants who were physically able to affect their own escape had
done so by that time. Further, the collapse of WTC 2 slowed the evacuation rate in WTC 1 due to the
significant quantity of dust and debris complicating the path from the bottom few stair landings out and
away from WTC 1 .
NIST NCSTAR 1-7, WTC Investigation 125
Chapter 10
WTC 1
WTC 2
10 15 20 25 30
09
35 40 45 50 55 60 65 70 75 80 85 90 95 100 105
Time Since WTC 1 Attack (min)
Figure 10-1. Reported evacuation time for survivors of WTC 1 and WTC 2.
§ 70%
=> 40%
30%
10%
20
40 60
Time from WTC 1 Attack (min)
80
WTC1
<~ollapsp
1 \\
►
\
\
\
\ WTC 1 WTC 2
\ Collapse
i WTC 1
! Attack
\ \
1 <
\ \
^ MITC 2 \
WTC 2 \ \.
Attack V ^V
4 ^ N
, -—- .,
100
Figure 10-2. Percentage of survivors remaining in the building for WTC 1 and WTC 2.
126
NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
While Figure 10-2 provides a comparison of overall rate of evacuation in WTC 1 and WTC 2, analyses
should also take into account that many occupants did not survive the attacks. Recasting Figure 10-2 to
include those who did not survive results in Figure 10-3, which shows the percentage of the total building
population remaining in the building over time. Note that the slopes of the two curves in Figure 1 0-3 are
not directly comparable since twice as many occupants perished in WTC 1 as did in WTC 2. When
occupants who were unable to evacuate are included, more than 80 percent of the occupants of WTC 1
survived and more than 90 percent of the occupants in WTC 2 survived.
=3
DQ
Q.
O
CD
Q.
CD
&
n.
<D
o
CD
Q.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
10
WTC 1
Collapse
I \x
\ ^S. WTC 1
M V
WTC 2
Collapse
►
iWTC 1 I \
j Attack V \
i 4 \wtc 2 ^>w
| WTC 2 V
| Attack -.
*«—«">►
20
30 40 50 60 70
Minutes From WTC 1 Attack
80
90
100
Figure 10-3. Percentage of occupants remaining in the building for WTC 1 and WTC 2.
Evacuation time, however, has two primary components which contribute to the total time: the delay time
prior to initiating evacuation and the time spent evacuating. Each of these components is discussed in
detail below, incorporating multiple regression analysis (causal modeling) of each.
10.1.1 Evacuation Initiation Delay Time in WTC 1 and WTC 2
Evacuation initiation delay time for occupants of WTC 1 and WTC 2 was discussed briefly in Chapter 6.
The analysis presented here explores whether statistically significant differences existed in the overall
evacuation initiation times for occupants in WTC 1 and WTC 2 and among occupants located in the high
(floor 77 and higher), middle (floors 43 to 76), and lower (basement to floor 42) regions of each tower
(for a total of six zones). Histograms of reported delay time based upon the telephone interview results
showed that each zone had a peak evacuation initiation around minute one or minute two, generally
decreasing over time (or skewed right). The general distribution for all six zones was a non-normal
distribution, approximated by a gamma or log-normal distribution.
NISTNCSTAR 1-7, WTC Investigation
127
Chapter 10
Note that due to the probable bias in the data from the high region in WTC 2 discussed previously
(occupants who delayed a length of time sufficient to prevent them from descending below the 78th floor
prior to 9:02:59 a.m. were disproportionately removed from the telephone interview sample), no further
consideration was given to comparing the results from the high zone in WTC 2 for the variable of
evacuation initiation delay time.
In order to calculate whether differences between the remaining five zones were statistically significant,
the reported data were transformed to better approximate a normal distribution and, thus, satisfy statistical
testing assumptions. A Box-Cox Power analysis revealed that all five zones had peak normal probability
plot correlation coefficients (of approximately 0.95 or better) when the Box-Cox power was
approximately zero, which indicated a log transformation. After the reported evacuation initiation delay
times were log-transformed, an analysis of variance (ANOVA) F-statistic was calculated for WTC 1 and
WTC 2.
Meaningful differences in the evacuation initiation delay time in WTC 1 among the three zones (high,
middle, low) existed, significant at the 1 percent level. Exploring the interzonal differences in WTC 1,
there was not a statistically significant difference in the reported evacuation initiation delay times between
the low zone and the middle zone. The high zone, however, exhibited significantly longer reported delay
times, when compared to both the low zone and the middle zone. The increased delay time in the high
zone compared to the low zone was significant at the 2 percent level, while the increased delay time in the
high zone compared to the middle zone was significant at the 1 percent level.
For WTC 2, evacuation initiation delay time was higher in the middle zone, compared to the lower zone;
the difference was significant at the 5 percent level. Again, comparing the high zone to either the middle
or lower zone in WTC 2 was not considered due to the bias in the upper zone delay time results.
In conclusion, the general trend in WTC 1 and WTC 2 was for evacuation initiation delay to increase with
building height, although there may be other explanations for this phenomenon. In WTC 1 , occupants
reported being trapped by debris and building damage, observing fire and smoke, and assisting injured
colleagues at a higher rate near the impact region (high) than was reported in the lower or middle zones.
The causal modeling in Section 10.2.1 explores this further through the independent variables 'floor' and
'environmental cues.'
Additionally, comparison of evacuation initiation delay times across buildings should be made with
considerable caution. In WTC 1, the cues which led occupants to decide to evacuate were substantively
different from those that likely influenced the decision process in WTC 2. In WTC 1, the building had
been attacked, stairwells were the only available route out of the building, and there were generally no
instructions delivered over the public address system. WTC 2 had not yet been attacked for over
90 percent of the occupants by the time they initiated evacuation. Thus, for less than 10 percent of the
WTC 2 survivors, their building had been attacked by the time they chose to initiate evacuation. In
addition, in WTC 2 stairwells and elevators were available for most people to choose to use, and multiple
(though conflicting) public address system announcements were made. Note, however, that many
occupants in WTC 2 did not hear any announcements, because they had evacuated the building prior to
the time the announcements were delivered.
128 NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
10.1.2 Time and Average Travel Speed in the Stairwells, WTC 1
The reported time spent in the stairwells for occupants of floor 10 to floor 91 in WTC 1, based on the
NIST telephone interviews (n=368), yielded a mean normalized travel speed of 1.3 floors per minute.
Normalized travel speed is defined as the total time from entering the stairwell until leaving the building,
divided by the number of floor that had to be descended. According to Figure 10-4, 25 percent of the
occupants traveled faster than 1.5 floors per minute, while 25 percent of the occupants traveled more
slowly than 0.9 floors per minute. The median normalized travel speed was 1.2 floors per minute. The
data include all interruptions to the evacuation process, including crowding, transfer floors, smoke, water,
switching stairwells, and resting. Assuming an average floor height of 3.7 m (12 ft), the distance along
the stair slope (including landings) would have been approximately 10 m (33 ft), yielding a movement
speed of approximately 0.2 m/s (0.65 ft/s) for the median occupant in WTC 1 while in the stairwells,
which is on the slow end of published scientific literature values for stairwell travel speeds. This is
understandable given the frequency of crowding and the significant number of obstacles to evacuation
reported by many occupants.
0.15
1 1.5 2
Floors per Minute
Figure 10-4. Stairwell travel speed in WTC 1 for all stairwells.
A "rule-of-thumb" for calculating evacuation flow rate is to assume that a standard size door at the
bottom of the stairwell is capable of discharging approximately one person per second. (Fruin 1987) 61 By
that logic, with three stairwells, the WTC system would have been capable of moving approximately
Fruin reported values of 40 - 60 persons per minute for free-swinging doors. The "rule-of-thumb" captures the high end of the
reported range.
NISTNCSTAR 1-7, WTC Investigation
129
Chapter 10
three people per second from the occupied floors to the Mezzanine or Concourse. In WTC 1, where
elevators were not operational for the duration of the evacuation, approximately 7,500 occupants used the
stairwells over a 100 minute period, yielding a flow rate of 1.3 people per second from the occupied
floors. Even discounting the final 27 minutes (after the collapse of WTC 2 when the flow rate dropped
significantly), the flow rate was 1.5 people per second (or 0.5 people per second per door), or about one-
half the "rule-of-thumb" flow rate. This is consistent with the previous observation that the stairwell
movement speed was slower than the published literature values would have predicted and that the slower
rate was to be expected given the number of obstacles to egress encountered by the evacuees and the total
travel distance required.
Impact of Wider Stairwell on Stairwell Evacuation Times
An analysis of the hypothesis that wider stairwells correspond to faster overall evacuation times is most
reliable when all other variables are held constant. In WTC 1 on September 11, 2001, however, many
other variables were not constant, complicating the comparison: (a) Stairwell B (56 in. wide) exited into
the Concourse, while Stairwells A and C (44 in.) exited to the Mezzanine where occupants typically
traversed to the escalator in order to descend to the Concourse; (b) Stairwell B only required one
(relatively short) horizontal transfer section, while Stairwells A and C required multiple, sometimes
lengthy (over 1 00 ft) horizontal transfers; (c) emergency response personnel preferentially used Stairwell
B to climb to higher floors; and (d) an occupant may have switched stairwells during egress, introducing a
significant uncertainty. Therefore, these four factors confound any conclusions regarding stairwell width
which may be drawn from the evacuation of the WTC towers. Respondents reported three pieces of
information critical to this analysis: number of floors they had to climb down, the total time spent in the
stairwells, and which stairwell they used. Each reported time was normalized (divided by) with the
number of floors descended in order to compare results independently of starting location. The
normalized times were then averaged over all occupants who reported using that stairwell.
In Stairwell B in WTC 1, the average occupant spent approximately 61 ± 38 s per floor, while in
Stairwells A and C (the narrower stairwells), the average occupant spent 53 ± 34 s per floor. The
uncertainty was calculated using a standard deviation. Unfortunately, the large uncertainty, relative to the
average, in the data collected from WTC 1, combined with the confounding conditions referred to
previously, do not allow differentiation of stairwell movement speeds for the wider stairwell compared to
the narrower stairwells.
10.2 FACTORS INFLUENCING TOTAL EVACUATION TIME
The telephone interview results permit rigorous statistical analysis of factors that played a role in the
evacuation process. In order to understand what factors influenced the overall evacuation time of the
average occupant in WTC 1 or WTC 2, two primary dependent variables were predicted: how long an
individual delayed initiating his or her evacuation, where initiation is defined as entering a stairwell or
elevator with the intention of exiting the building; and how long an individual spent traversing the
stairwells. Elevator travel time was not considered in this analysis. The sum of these two elements of the
evacuation process is the total evacuation time. Multivariate regression modeling was utilized to establish
factors which contributed to overall evacuation time. A more detailed discussion of both causal models,
including methods, equations, and significance testing can be found in Appendix C: Causal Modeling.
This chapter will only present the basic model structure and findings.
1 30 NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
10.2.1 Predicting Evacuation Delay in World Trade Center
The analysis of evacuation behavior consisted of two stages. The first stage, which focused on how long
it took for people to begin evacuating, determined the factors (variables) and social processes (the major
paths of causal links between variables) that influenced people delaying the initiation of their evacuation
out of WTC 1 and WTC 2 on September 11, 2001. Evacuation delay was defined as the number of
minutes that passed from when a person first became aware that something was wrong until he or she
began evacuating.
Model Description
The model used to predict important factors in evacuation delay in the towers used variables that
preliminary analyses suggested were salient and closely followed general evacuation theory from the
social and fire sciences The model is diagrammatically illustrated in Figure 10-5. It can be described as
follows: (1) delay in evacuation initiation (X 7 , the dependent variable in this analysis) was a direct
consequence of environmental cues (Xi), the floor on which occupants were located (X 2 ), obtaining
information without seeking it (X 3 ), perceived risk (X 4 ), seeking additional information (X 5 ), and taking
pre-evacuation actions (X 6 ); (2) taking pre-evacuation actions was a direct consequence of environmental
cues, floor, obtaining information without seeking it, perceived risk, and seeking additional information;
(3) seeking additional information was a direct consequence of environmental cues, floor, obtaining
information that was not sought, and perceived risk; and, finally, (4) perceived risk was a direct
consequence of environmental cues, floor, and obtaining information without seeking it. This model was
parsimonious (in other words, the model was constructed of no more or fewer components than were
necessary), and consistent with theory that stems from research on existing evacuation and risk
communication research, and the model accurately reflected findings from preliminary analyses on the
many variables that could have impacted evacuation delay. With this conceptual model, a set of equations
for evacuation delay, pre-evacuation actions, seeking additional information, and perceived risk that were
assumed to depend linearly on the factors above were solved simultaneously to obtain the relative
importance of each factor in predicting the four primary variables.
Results
The estimated parameters of the models for WTC 1 and WTC 2 revealed that the model had a very high
degree of success in explaining evacuation initiation delay, pre-evacuation actions, seeking information,
and perceived risk in both towers. The adjusted explained variance (R 2 ) for perceived risk was 55 and
60 percent in WTC 1 and WTC 2, respectively. The adjusted explained variance for sought information
was 25 percent in both towers. Respectively, the adjusted R 2 for pre-evacuation actions was 68 and
69 percent for WTC 1 and WTC 2. Finally, the adjusted R 2 for delay in evacuation for WTC 1 and
WTC 2 was, respectively, 49 and 56 percent. These were extraordinarily high levels of adjusted explained
variance to observe in a study of human evacuation. The combined amount of variance in the dependent
variable, evacuation thereby, established the strong predictive power of the models for both towers. All of
the equations in the model for WTC 1 and WTC 2 were statistically significant at the 0.001 level or
higher, indicating a high degree of confidence in the explanatory power of the model.
NIST NCSTAR 1-7, WTC Investigation 1 3 1
Chapter 10
/
/
Perceived
Risk
Pre-Evacution
Actions
fc.)
Environmental
Cues
/
e 7
Delay
Initiating
Evacuation
(x,)
Obtained
Information
M
Sought
Information
Figure 10-5. Model of evacuation delay.
Direct Effects
With a few exceptions, similar findings regarding evacuation delay emerged in both WTC 1 and WTC 2.
Repeatability of findings lends validity to the conclusions that can be drawn from the analyses. The
findings are presented below. First, each of the four equations in both models for both towers was
considered; next, the models were interpreted as a whole so that the most significant paths of influence for
each tower could be distinguished.
Predicting Perceived Risk
The findings that emerged regarding predicting the risk that people perceived were virtually identical
across the two towers. 'Perceived' risk was considered more important than objective or actual risk
because identical observations may be interpreted differently by individuals, and because people act on
the basis of perceptions, or what they believe to be true. The R 2 for perceived risk was 55 percent in
WTC 1 and 60 percent in WTC 2. In WTC 1, both environmental cues and floor had strong and similar
impacts on predicting perceived risk, while obtained information had a weaker but statistically significant
132
NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
impact. In WTC 2, once again, both environmental cues and floor had strong impacts on predicting
perceived risk, and, once again, obtained information had a weaker but statistically significant impact.
These findings suggest that the risk that people perceived before they began their evacuation increased
largely as a function of their starting floor (distance to safety) and being exposed to environmental cues.
Clearly, information had a lesser impact on risk perception than did the two more salient variables of
experiencing environmental cues and floor height. The only difference in findings between the towers
was that, in WTC 2, floor height was by far the strongest predictor of perceived risk, while in WTC 1,
both floor and environmental cues were equal predictors. In other words, those who were on higher
floors felt more at risk and moved more rapidly to evacuate. 62
Predicting Seeking Information
The variable 'sought information' may include verification of the nature of the attack, instructions or
directions, or other information deemed relevant prior to evacuation initiation. Once again, the findings
that emerged for predicting seeking information were almost identical across the towers. Explained
variance (R 2 ) for seeking information was 25 percent in both WTC 1 and WTC 2. In WTC 1,
environmental cues and floor both had the strongest and identical impacts on seeking information;
obtained information had no statistically significant impact on seeking information; and perceived risk
had a slight impact on seeking information. In WTC 2, environmental cues had the strongest impact on
seeking information; the impacts of floor and obtained information were not significant; and perceived
risk had a slight impact on seeking information.
Seeking information in times of rapid onset emergencies is a typical human response, since people need
to interpret and make sense out of an event before they act on it. The finding that the variable of
environmental cues was the strongest predictor of seeking additional information is consistent with this
theoretical finding about "milling" from past research. Obtained information had no impact on seeking
information in either tower. This was likely because information to make sense out of the event had
already been obtained. Perceived risk had a similar effect on seeking information — albeit lesser of an
effect than environmental cues — in both towers. Interestingly, floor height (or distance from the exit) had
a significant effect on seeking information in WTC 1, but not in WTC 2. Evacuation theory would
predict that this effect would be present; it was present in the tower that was struck first, and it was not
present in the tower struck second.
Predicting Pre-Evacuation Actions
Pre-evacuation actions included activities such as making telephone calls, gathering belongings, or
performing any other tasks deemed necessary prior to evacuation initiation. The explained variance (R 2 )
for taking pre-evacuation actions in WTC 1 was 68 percent, and it was 69 percent in WTC 2. In both
towers, the strongest predictor of taking pre-evacuation actions was floor. Environmental cues was also
predictive of pre-evacuation actions. Once again, observing cues indicating that one is at risk and being
high in the building with a longer path to safety emerged as strong predictors. Obtaining information had
virtually no impact on pre-evacuation actions in either tower. Seeking information impacted pre-
evacuation action likely because the information obtained supported the need to evacuate and, hence,
Recall, however, that individuals who were not motivated to initiate evacuation on higher floors in WTC 2 were more likely to
be trapped above the aircraft impact region when WTC 2 was attacked.
NIST NCSTAR 1-7, WTC Investigation 1 33
Chapter 10
related to getting ready to leave. Finally, the impact of perceived risk on taking pre-evacuation actions
was higher in WTC 2 than in WTC 1 .
Predicting Delay in Evacuation
Explained variance (R 2 ) in evacuation delay was 49 percent in WTC 1 and 56 percent in WTC 2. The
greatest predictor of evacuation delay in both towers was taking pre-evacuation actions. Obviously, doing
anything before initiating evacuation — including preparation to leave — delayed departure. Setting this
factor aside, some clear differences emerged between the two towers in terms of the relative impacts of
the remaining variables in the model. Perceived risk had no direct effect on evacuation initiation delay.
This finding is consistent with general evacuation theory where perceived risk's impact on actual
behavior is indirect through other factors. The three factors with the strongest direct effects on evacuation
delay were the same in both towers. These were environmental cues, floor, and obtained information. In
both towers, floor's effect was negative, that is, the more floors one was from the exit, the quicker people
were to initiate their evacuation. Environmental cues and information that was received passively both
increased delay in the initiation of evacuation. Finally, seeking additional information had a minimal
impact on evacuation delay.
Paths of Greatest Influence and Conclusions
Although each of the aforementioned findings are interesting in and of themselves, perhaps the most
important findings that we can offer are those that emerge when all of the individual findings offered
above are brought together and viewed at the same time in the context of the entire model.
Bias
As discussed previously, any conclusions about evacuation initiation delay time in WTC 2 should
consider the impact of disproportional decedent location, particularly as a source of the disproportionality
may be highly correlated to the variable of interest, evacuation delay. In other words, those who
exhibited long delay times in one region of the building were unable to be interviewed, thus artificially
shortening the average delay time for one-third of the building. In the causal modeling, this would affect
the relationship between 'floor' and 'delay initiating evacuation,' likely tending towards zero a slightly
negative estimate (-0.19) of the beta value between the two variables. As floor was not a primary path
which directly predicted evacuation initiation delay in WTC 2, the impact of this bias was considered
secondary. The effect of this bias as it worked through other variables was not considered.
WTC1
Although there were other factors that had some lesser impacts on influencing what people did, the paths
of causal influence that defined the main process that led to delay in the evacuation of WTC 1 on
September 11, 2001 follows.
Environmental cues (information from the physical environment that something was terribly wrong) and
floor (increased distance to safety) caused people to set out to find additional information, most likely
information about what was going on and what they should do. Next, the act of seeking additional
information, that is, "milling" about to make sense out of the situation, led people to take actions to
1 34 NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
prepare to evacuate. Finally, taking those actions to prepare to evacuate delayed the initiation of actually
evacuating.
In addition to this four step causal process, environmental cues and floor also had indirect impacts on
evacuation delay as follows. Both factors increased the odds of seeking information and both factors
increased the chances that people would take pre- evacuation actions prior to evacuating. Both factors also
had direct impacts on actual evacuation delay. Environmental cues increased delay while floor decreased
delay.
WTC 2
Although there were other factors that had lesser influence on what people did, the paths of causal
influence that defined the main process that led people to delay in the evacuation of WTC 2 on
September 11, 2001, were identical to WTC 1 with one decided difference.
Environmental cues (information from the physical environment that something was terribly wrong) and
floor (increased distance to safety) predicted perceived risk. Environmental cues, floor, and perceived
risk caused people to set out to find additional information. Next, the act of seeking additional
information, that is "milling" about to make sense out of the situation, and perceived risk both led people
to take actions to prepare to evacuate. Finally, taking those actions to prepare to evacuate delayed the
initiation of actually evacuating.
In addition to this five step causal process, environmental cues and floor also had indirect impacts on
evacuation delay. Both factors increased the odds of seeking information and both factors increased the
chances that people would take pre-evacuation actions prior to evacuating. Both factors also had direct
impacts on actual evacuation delay. Environmental cues increased delay, while floor decreased delay.
10.2.2 Predicting Normalized Stairwell Evacuation Time in WTC 1 on
September 11, 2001
The second component of total evacuation time was the time spent in the stairwells. This analysis
determined the factors and social processes that influenced the normalized stairwell evacuation time per
story of stairs for the people who evacuated out of WTC 1 on September 11, 2001. WTC 2 was excluded
from this analysis because evacuees used stairs, elevators, and/or a combination of both for their
evacuation and could not be separated for the analysis. Evacuation time was defined as the average
number of seconds per story of stairs that it took people from the time they entered a stairwell until they
completed their evacuation out of the building. The model used to predict important factors in stairwell
evacuation time again used variables that preliminary analyses and general evacuation theory suggested
as salient, and is illustrated in Figure 10-6. It can be described as (1) normalized stairwell evacuation time
is a direct consequence of floor, evacuation decision delay, environmental cues, emergency responders,
crowding, and evacuation interruption; (2) evacuation interruption is a direct consequence of floor,
evacuation decision delay, environmental cues, emergency responders, and crowding; (3) crowding is a
direct consequence of floor, evacuation decision delay, environmental cues, and emergency responders;
NIST NCSTAR 1-7, WTC Investigation 1 35
Chapter 10
Evacuation
Decision
Delay
Emergency
Responders
Evacuation
Interruption
Normal
Stairwell
Evacuation
Time
Environmental
Cues
x,
Crowding
Figure 10-6. Causal model for predicting normalized stairwell evacuation time.
(4) emergency responders is a direct consequence of floor, evacuation decision delay, and environmental
cues; (5) environmental cues is a direct consequence of floor and evacuation decision delay; and, finally,
(6) evacuation decision delay is a direct consequence of floor. This model is parsimonious, consistent
with the high rise fire evacuations theory, and it well-represented the positive findings of our preliminary
analyses of the many variables that could have impacted normalized stairwell evacuation time.
Results
The estimated parameters of the model for WTC 1 revealed that the model had a very high degree of
success in explaining normalized stairwell evacuation time. The adjusted explained variance (R 2 ) for
normalized stairwell evacuation time was 44 percent, 1 1 percent for evacuation interruption, 72 percent
for crowding, 57 percent for emergency responders, 79 percent for environmental cues, and 34 percent for
evacuation initiation delay. With the exception of evacuation interruption, these are extraordinarily high
levels of adjusted explained variance to observe in a study of human evacuation; and, these R 2 's, thereby,
establish the strong predictive power of the model. All of the equations in the model were statistically
significant at the 0.001 level or better.
Direct Effects in the Model
We first consider each of the factors individually, and then the model is interpreted as a whole so that the
most significant paths of influence can be distinguished.
136
NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
Predicting Evacuation Initiation Delay
Evacuation initiation delay is the total time from first awareness that something was wrong on
September 11, 2001, until entering a stairwell to evacuate. The findings that emerged regarding
predicting delay in the initiation of evacuation from floor were that the R 2 was 34 percent. This
relationship is already discussed in detail in the previous section.
Predicting Environmental Cues
Environmental cues were the visual, auditory, or other sensory perceptions that indicated danger on
September 11, 2001. Explained variance (R 2 ) for observing environmental cues was 79 percent, and the
equation was statistically significant at the 0.001 level. Floor had a very strong direct impact on observing
environmental cues, and it was significant at the .001 level. The effect of delay in the initiation of
evacuation was smaller, but still statistically significant at the 0.001 level. It would appear that the longer
a person took to begin their evacuation, the more the physical impacts of the event grew and the more
likely people were to experience them.
Predicting Emergency Responders
Emergency responder was denoted as either observed or not observed. The explained variance for
predicting encountering emergency responders was 57 percent, and the equation was statistically
significant at the .001 level. Experiencing environmental cues and floor both predicted encountering
emergency responders. This makes sense when one considers that emergency responders would be most
likely to go to areas experiencing the impacts that would also yield environmental cues, and the higher
one was in the tower, the more stairwells one had to traverse, the longer the traversal time overall, and,
thus, the greater the odds of encountering emergency responders.
Predicting Crowding
Stairwell crowding was denoted when a respondent indicated that the stairwell was crowded to the extent
that it impacted progress down the stairs. The explained variance for perceived crowding on the
evacuation stairwells was 72 percent. Perceived crowding largely increased as a result of environmental
cues and encountering emergency responders.
Predicting Evacuation Interruption
Evacuation interruption was denoted by respondents who indicated that they left the stairwell or stopped
in the stairwell by choice. The explained variance for interrupting evacuation was only 1 1 percent. Even
though this equation was statically significant at the .001 level, none of the five predictor variables in the
equation had a statistically significant impact on evacuation interruption. At best, these findings can be
interpreted to mean that there was a slight but not statically significant tendency for people to interrupt
their evacuation if they had more rather than fewer floors to traverse to safety, and if they encountered
environmental cues (perhaps obstacles) in the process of evacuation.
NIST NCSTAR 1-7, WTC Investigation 1 37
Chapter 10
Predicting Normalized Stairwell Evacuation Time
Explained variance in predicting normalized stairwell evacuation time was 44 percent, and the equation
was statistically significant at the .001 level. Only two factors in the equation were statistically
significant, both at the .001 level. These were environmental cues and evacuation interruption. Clearly,
the single factor that had the biggest impact on increasing the amount of time people spent, on average,
per stairwell was environmental cues. The only other factor that had a significant impact was interrupting
evacuation, obviously, because stopping egress would increase the amount of time needed to complete
evacuation.
Paths of Greatest Influence and Conclusions
Although each of the aforementioned findings are interesting in and of their own right, the most important
findings that we can offer are those that emerge when all of the above findings are brought together and
viewed at the same time in the context of the model as a whole. Thus, the main process that led to
increased normalized stairwell evacuation time in the evacuation of WTC 1 on September 1 1th was
straightforward and clear. Floor (increased distance to safety) substantially increased the odds that people
would encounter environmental cues. Floor also increased delay in starting evacuation (this relationship is
elaborated upon in much detail previously), which, in turn, also increased the chances that people would
encounter environmental cues. But it was encountering environmental cues that had a large and direct
effect on increasing the amount of time that people spent, on average, to traverse their evacuation
stairwell. In addition to this multi-step process with environmental cues as the key predicting variable,
interrupting the process of evacuation for any reason also increased the amount of time, on average, that
people used to descend their evacuation stairwell.
10.3 ISSUES THAT IMPACTED OCCUPANT EVACUATION
While the multivariate regression analysis of the telephone data has extraordinary power to distill the
telephone interview data to the salient features, the analysis was necessarily limited by the constraints of
the number and type of questions contained in the telephone interviews. Face-to-face interviews and
focus groups provided a more detailed understanding than can be achieved by exclusive use of the
telephone interviews. This section provides a discussion of some of the issues identified from both the
telephone survey and face-to-face interviews that impacted the ability of occupants to successfully
evacuate the two towers on September 1 1, 2001, including both those issues that aided evacuation and
those that made evacuation more difficult.
10.3.1 Environmental Cues and Information
In any emergency, occupants obtain information about the event by passively receiving it and actively by
seeking it out through various means. Information may have led occupants to evacuate or to delay
evacuation to gather additional information or to perform additional tasks prior to beginning evacuation.
As soon as WTC 1 was struck, most occupants' first reaction was to look out the window for information.
These first environmental cues from the building (hearing the explosion, feeling the building shake, etc),
coupled with the reaction of looking out the window (and seeing the fireball and smoke in WTC 1 , the
debris falling outside of the window, or even the plane hitting the tower), was enough to signal that
138 NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
something was wrong for many in the towers. This first set of information caused some to begin the
evacuation process immediately.
An occupant in WTC 1 from a floor in the 70s reported immediate
evacuation initiation: "The building was swaying more than normal and
I slammed into the cabinets. I got up and headed for the staircase."
Interview 1000574 (NIST 2004)
Also, an occupant in WTC 2 from a floor in the 80s also noted
evacuating immediately. "Somebody in my area jumped up and said
there is a bomb. At this stage, everybody else took the stairs. I went
directly to the Stairwell C. I didn't think I was in danger at this point."
Interview 1000568 (NIST 2004)
In both towers, many delayed evacuation to find out more information before entering a stairwell or
elevator. This may have involved talking with coworkers, making and receiving phone calls, sending and
receiving emails, searching the internet, and watching TV.
According to the telephone interviews, 1 1 percent of the occupants in WTC 1 and 21 percent of the
occupants in WTC 2 received information related to the event through various means. Of the occupants
who received information, 57 percent in WTC 1 and 65 percent in WTC 2 received information about
what happened. A majority of the occupants in WTC 2 (54 percent) received information about what
happened to WTC 1 through the public address announcement. Many of them, already in the process of
evacuating, were faced with the decision of whether to return to their offices or continue their evacuation.
While on their floor, occupants were receiving information about what was going on in the buildings from
each other and through a variety of technological means. Technology such as landline phones, cell
phones, blackberries, email, internet, and TV was used inside the building to receive such information.
An occupant from a floor in the 80s of WTC 1 received a phone call
from a co-worker. "A co-worker was calling from home. She said that
'a plane hit the building and you better get out of there.'" Interview
8000005 (NIST 2004) This information caused the occupant to begin
their evacuation process.
However, the situations in the towers were different before WTC 2 was hit. The information that was
received by the occupants in both buildings was that only WTC 1 had been hit by an aircraft. This
message produced a variety of responses from occupants in WTC 2.
After hearing the public address system announcement, an occupant from
a floor in the 50s in WTC 2 was calling an out-of-town office to let them
know they were evacuating the building. "I got a hold of them and they
told me that a plane hit the building [WTC 1] and to get out of there."
The respondent hung up and continued to call loved ones and answer
other phone calls until his or her own building was hit by an aircraft.
Interview 1001666 (NIST 2004)
An occupant in the 40s from WTC 2 was watching TV before WTC 2
was hit. "I saw what happened on CNN and was informed that a plane
NIST NCSTAR 1-7, WTC Investigation 1 39
Chapter 10
hit Tower 1. I entered the emergency stairwell A and proceeded down.
Interview 1000867 (NIST 2004)
Even after entering the stairs, the evacuation flow continued through the use of technology and
information sharing among the occupants inside the building. Although they had already entered the
stairs, additional information (aside from the public address system announcement in WTC 2 telling
occupants to return to their floor) sometimes increased the occupants' motivation to evacuate the
building, as shown in the following quotes from WTC 1 occupants.
An occupant from a floor in the 50s in WTC 1 was traveling in a
stairwell when he or she heard another occupant say "that another plane
hit the building. [The person] got this information from a cell phone and
we speeded up our evacuation." This respondent went on to indicate that
"cell phones were working and people were trying to figure out what was
going on." Interview 1000071 (NIST 2004)
While in Stairwell B, an occupant from a floor in the 70s in WTC 1
"heard a fire department radio say the words '. . . Structural Instability . .
. '" This radio message was relayed at the time of the collapse of
WTC 2. The respondent went on to say that they "stopped because there
was no movement in the staircase, but then became frightened for the
first time. [I] began to focus more on getting out." Interview 1000118
(NIST 2004)
An occupant in WTC 2 also experienced the same knowledge about his
building during evacuation. "The phone was ringing, [so] I answered the
phone on the 3rd floor. I listened to my daughter tell me that both
buildings were hit [by a] commercial airliner, and that it was a terrorist
attack. I told my boss what I learned from my daughter, so we would
both have more determination to get out." Interview 1000003
(NIST 2004)
When people were given new information about the event from outside of the building, they shared that
information freely with those around them. People provided information about the phenomena that they
witnessed themselves (at first awareness) or information that they had received from another source, such
as from someone outside of the building. As evidenced in face-to-face interviews, the behavior of sharing
information was frequent, regardless of whether the individuals involved were strangers.
For example, there was an occupant in WTC 1, who originated on a floor
in the 40s, whose phone rang while in the stairwell: "I answered the
phone and my [spouse] was telling me that another plane had struck
Tower 2. [The spouse] said to get out of the building as fast as you can
and that it is a terrorist attack. [The spouse] also told me about the plane
hitting the Pentagon. [The spouse] mentioned that there was another
plane in the air but [the spouse] wasn't sure where it was. I told others
around me what had happened, in order to spread the available
information." Interview 1000572 (NIST 2004)
In addition to passively receiving information from inside the buildings, 28 percent of the survivors in
WTC 1 and 29 percent of the survivors in WTC 2 actively sought out information related to the event
1 40 NIST NCSTAR 1 - 7, WTC Investigation
Discussion and Analysis
through various means. In the face-to-face data, this includes nearly everyone when the occupants who
looked out the window for information are included. However, the telephone interview data shows that
approximately 40 percent of these who sought out information about the event in both towers were not
successful in finding the information they were searching for.
For example, an occupant from a floor in the 60s of WTC 1 tried making
several phone calls after the building was hit, but was unsuccessful.
"The building started to sway and everything started shaking. I knew
that there was something [wrong]. I ran to my desk and made a couple
of phone calls. I dialed about five times trying to reach my [spouse]. I
also called my sisters to find out more information." Interview 1000733
(NIST 2004)
Occupants also attempted to seek information about what was going on in the buildings through a variety
of technological means. This technology included landline phones, cell phones, blackberries, email,
internet, and television broadcasts inside the building.
An occupant in the 70s performed a variety of information seeking
activities before deciding to leave WTC 1 . "I was curious and wanted to
get information on what happened. I tried to find information on the
internet, [because] that was the fastest way of getting information.
[Also], I knew my mother had TV access. I called my mother to get
information on what she had seen so far on TV." This occupant also
phoned a brother-in-law for more information until the occupant
witnessed the plane impact WTC 2. Then, the occupant began to
evacuate. Interview 1000583 (NIST 2004)
Also, an occupant from a floor in the 100s in WTC 2 noted that "there
were a lot of people trying to find out what was going on by using cell
phones." Interview 1000563 (NIST 2004)
Emergency responders also confirmed what happened to inquiring occupants inside the stairwell, but then
reassured them that everything would be okay. Obtaining information from emergency officials inside
the building was not frequently reported. In return, occupants would provide information to the
firefighters about where certain people were waiting for rescue assistance.
Overall, the information flow and its impact on occupants' evacuation were dependent upon many factors,
including the tower in which the occupant was located and the time the information was received during
the emergency. Occupants in WTC 2 were sometimes affected differently to news that only WTC 1 had
been hit by aircraft, when compared to occupants in WTC 1. Information was received and sought out at
various times throughout an occupants' evacuation and from different sources. However, many
occupants in both towers reacted to the first set of information provided by the plane crash into WTC 1,
either by initiating evacuation or by beginning pre-evacuation activities.
10.3.2 Building Alarm Systems
Fire alarms have been a hallmark of fire protection for more than a century, with systems proposed in
New York City as early as 1847 (Bukowski and Moore 2003). Typically, the systems serve dual purposes,
NIST NCSTAR 1-7, WTC Investigation 141
Chapter 10
occupant notification and transmission of information about the fire to emergency responders. On
September 11, 2001, with significant damage to the fire alarm system in the towers, only a small
percentage of occupants in both towers (14 percent in WTC 1 and 10 percent in WTC 2) noted hearing a
fire alarm during different stages of their evacuation. Many occupants in WTC 1 had already begun to
evacuate their building due to a variety of building observations from the plane impact (less than
5 percent of occupants in WTC 1 said that their decision to evacuate was based on the fire alarm going
off). The same is true for occupants in WTC 2 who heard alarms after their tower was hit, since a large
majority had already begun evacuation before or when their tower was hit (less than 1 percent said their
decision to evacuate was based on hearing a fire alarm).
Some occupants of WTC 2 believed that the emergency was limited to WTC 1 and were unclear if their
evacuation was necessary. Also, occupants in WTC 2 were faced with a public address system
announcement informing them to return to their offices. For example, a respondent in WTC 2 heard the
announcement that the building was secure first, followed by the fire alarm.
"At the 25th floor, I heard someone on the loudspeaker who said that 1
World Trade was hit but that 2 World Trade was OK and that we should
all go back to our office. After I heard that announcement, I heard the
general alarm which meant to me to leave the building even quicker."
Interview 1000740 (NIST 2004)
And, in response to this, the occupant kept walking down the stairs to leave the building.
In both towers, occupants reported hearing fire alarms at various floors throughout the tower. For those
occupants in WTC 1 and WTC 2 who heard fire alarms on their floors, the fire alarm was not frequently
nor specifically mentioned as the reason for evacuating. Instead, the top three reasons for occupants to
begin evacuation from either towers were the plane hitting WTC 1 , being told to evacuate, or feeling
afraid or in danger (in no particular order for each tower).
10.3.3 Public Address Announcements
Public Address announcements were not mentioned in the face-to-face interviews as being heard after
each building was hit by aircraft. However, from the telephone survey, 14 percent of the survivors in
WTC 1 reported receiving information from a public address system announcement - which suggests that
the occupants received an announcement after their tower was struck. There is also information from the
9-1-1 tapes that a second announcement was heard in WTC 2 by people near and above the floors of
impact after the building was hit, although NIST found no evidence from face-to-face interviews that
survivors below the floors of impact heard this announcement.
A building-wide public address announcement was made to the occupants in WTC 2 approximately three
minutes before the aircraft hit their building. Based upon analysis of many face-to-face interviews and
published accounts, the first announcement provided to occupants in WTC 2, before their tower was hit,
stated the following:
There is a fire condition in WTC 1 , WTC 2 is secure/safe, please return
to your offices.
142 NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
Fifty-four percent of the occupants in WTC 2 reported receiving information from the public address
system. Since 40 percent of WTC 2 survivors had evacuated before WTC 2 was attacked, the
overwhelming majority of occupants still in the building heard the 9:00 a.m. announcement. While some
occupants interpreted the first announcement as a suggested course of action, other interpreted it as an
instruction.
Approximately two minutes later, at 9:02 a.m., one minute before WTC 2 was attacked, a second
announcement was made, contradicting the first announcement. Recorded in the background of an
answering machine receiving a call from an occupant calling home from floor 98, the announcement
indicated that it was now permissible to initiate an evacuation:
"May I have your attention please. The situation is in Building 1.
However, if conditions on your floor warrant, you may wish to start an
orderly evacuation." Interview 3000001 (NIST 2004)
The 9:02 a.m. announcement was noted by one occupant catalogued in the collection of published
accounts (Fahy and Proulx 2003), however, the large discrepancy in the percentage of occupants
reporting this announcement when compared to the percentage who reported the 9:00 a.m. announcement
remains unresolved.
The majority of occupants in WTC 2 began their evacuation before the first announcement was made.
Some of them never heard the announcement (due to evacuating via elevator, starting in the lobby, or
possibly evacuating from a lower floor before the announcement was given) and some heard the
announcement from either the stairs or a lower floor than their original starting floor. It is known that
occupants listened to the announcement from the stairs, left the stairs to hear the announcement more
clearly, or were led out of the stairs minutes before the announcement on skylobby floors by security
personnel. Those who heard the announcement and had begun the evacuation process were faced with a
decision on whether to comply with the instruction and return to their offices or continue their evacuation.
The face-to-face interview data shows that while many people did not follow the instruction/suggestion in
the announcement, others chose to return to their office. Each individual's decision was affected by a
number of factors at the time of the announcement, including:
• their perception of danger and ability to evacuate:
"While I was on the steps, I heard a public address system announcement
that said the building was secure and to return to our offices. It made me
angry to hear this announcement because I felt we were in danger."
Interview 1000048 (NIST 2004) This occupant, who had a medical
disability, was helped down the stairs by a fellow occupant, until they
took an elevator to the ground floor.
• the actions or statements of others in the immediate area (group behavior):
An occupant from a floor in the 70s, who used both an elevator and a
stairwell, heard an announcement around floor 20: "[I heard] a PA
(public address) announcement that told people to return to their offices.
I shouted to people to continue down the steps; to keep the flow of traffic
going—so that people wouldn't be going past me. I intentionally lied to
NIST NCSTAR 1-7, WTC Investigation 143
Chapter 10
people. I heard the PA [public address] announcement tell people to
return to their offices, but I thought it was better for them to evacuate the
building and I knew that if the people didn't turn around and try to move
past me that I could evacuate faster." Interview 1000024 (NIST 2004)
An occupant from a floor in the 60s in WTC 2: "I heard a public address
system announcement that said to stay where we were—that our building
was secure and the problem was with the other building, but the fire
warden said to ignore the message and to leave the building anyway."
Interview 1000050 (NIST 2004)
An occupant from a floor in the 90s heard an announcement around floor
78 in WTC 2: "We heard an announcement that said 'the building is safe;
you can go back.' I continued walking down the stairs, to keep moving,
to not hold up the others that were behind me." Interview 1000070
(NIST 2004)
• desire to follow instructions:
An occupant from a floor in the 30s in WTC 2: "The PA (public address)
announcement informed us, '. . . You may return to your office.' I exited
the stairwell on the [a floor in the 30s] and proceeded to the elevator
lobby on that floor. I entered the elevator and took it to [a floor in the
30s] to return to my office as instructed. I [then] called my [spouse] at
about 9:03 a.m. to tell [the spouse] that everything was ok - but my
conversation was cut short when the second aircraft struck Tower 2."
Interview 1000049 (NIST 2004)
• and even serendipity:
An occupant from a floor in the 10s in WTC 2: "I heard an
announcement saying that the problem is in Building One and for us to
stay where we are or to return to our floors. I stepped off of the stair into
the 9th floor to take the elevator to return to my floor. I was distracted
because of conversation and did not realize that the elevator was going
down until it started moving." Interview 1000922 (NIST 2004) This
occupant did not return to their office, but instead, evacuated the building
after exiting the elevator.
Occupants who were led out onto skylobby floors also hesitated after hearing the announcement. They
engaged in milling with other coworkers in an attempt to decide what to do. Minutes later, WTC 2 was
struck, which caused them to initiate their evacuation.
Other occupants who did not begin the evacuation process were still located on their floor when the PA
(public address) announcement was made. The announcement helped to reinforce their original goal of
remaining on their floor for some time. Most of these occupants listened to the announcement and
remained on their floor.
"I heard an announcement on the intercom that told us to stay. I decided
to stay to avoid getting in more danger. I heard the phones ringing off
144 NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
the hook and my boss and I were going from phone to phone answering
them. This was to reassure the families that everyone was ok."
Interview 1000524 (NIST 2004)
Regardless of where occupants were located in the building and whether they had decided to begin the
evacuation process, their final evacuation decision was prompted by the impact of the aircraft into their
building, WTC 2, at 9:02:59 a.m.
10.3.4 9-1-1 System
NIST was given access to all emergency calls to the New York City 9-1-1 system related to the World
Trade Center attacks on September 11, 2001 from 8:45 a.m. until shortly after 10:30 a.m. The 9-1-1
system was flooded with calls from 8:46:30 a.m., when WTC 1 was attacked, until roughly 8:55 a.m., and
experienced a second surge in call volume at 9:03 a.m., when WTC 2 was attacked, which continued for
several minutes. The majority of calls were eyewitnesses reporting that they had observed something
crashing into the World Trade Center or reporting that the World Trade Center was on fire. Initial reports
included descriptions of small planes, helicopters, large (commercial or military) aircraft, as well as a
bomb or a missile. A few callers sought advice about whether they should evacuate their nearby
buildings or information about what was going on. Other callers reported observations of suspicious
activities or people. After several minutes, the first WTC occupants began to call 9-1-1 seeking advice or
rescue assistance. Occupants continued to use the 9-1-1 system, both as a resource for information and an
outlet for reporting their situation, until around 10:00 a.m., when WTC 2 collapsed.
During the time period between aircraft impacts, advice from 9-1-1 operators to occupants experiencing
smoky conditions was largely to shelter in place and await rescue from emergency personnel.
One of the common questions asked by trapped occupants of the 9-1-1 operators was whether to break
windows in order to obtain fresh air. The advice given to occupants varied by operator. Some operators
encouraged the occupants to assess their own unique situation and determine whether breaking a window
would help, sometimes warning the occupants that breaking a window may only serve to introduce more
smoke from the outside than it would relieve from the inside. Other operators simply advised occupants
not to break windows. The 9-1-1 operators, largely acting without complete knowledge of an evolving
and traumatic situation, were forced to strike a balance between efficiently logging incoming calls,
compassionately counseling sometimes desperate occupants, and disseminating relevant event
information.
Another common question from occupants to 9-1-1 operators was whether they should evacuate, and
further, whether that should be upward or downward (in other words, where was the impact region
relative to their location). Some 9-1-1 operators did not initially offer advice, routinely telling the
occupants to defer to the instructions they were (presumably) receiving from building authorities. Other
operators consistently advised occupants to stay in place, rescue was on the way. At least one occupant
below the floor of impact in WTC 2 was repeatedly advised to await rescue by emergency personnel who
were advised of the occupant's whereabouts. Sometime after 9:02:59 a.m., some operators began to
encourage the occupants to evacuate, if possible. Information about the location or magnitude of the
impact in either WTC 1 or WTC 2 was not generally communicated to trapped occupants.
NIST NCSTAR 1-7, WTC Investigation 145
Chapter 10
Operators routinely collected from callers specific location information, which proved useful during the
NIST investigation for establishing the time-evolving building conditions (fire, smoke, damage, number
of people), including: building, floor, corner (northwest corner, for example), or office number.
Additionally, callers would sometimes relay individual or group observations about the tenability or
status of the egress path (i.e., access to stairwells or elevators). For these reasons, the 9-1-1 recordings
proved to be a valuable record of the conditions above the impact floors for which there would have been
no other method to discover.
10.3.5 Previous Evacuation Experience
Using prior evacuation experience to guide future evacuation decisions, may or may not produce better
outcomes. Recall from Section 4.3 that 16 percent of survivors on September 11, 2001 were also present
during the 1993 bombing. A survivor from the 70s in WTC 1 described:
"My response to the '93 bombing wasn't sufficient. I realized there was
no real purpose in sticking around. I was going to get out as quickly as I
could." Interview 1000525 (NIST 2004)
Another survivor from a floor in the 70s in WTC 1, however, used their 1993 evacuation experience to
justify delaying their evacuation:
"I was there in 1993, and [this time] I wanted to wait for some directions
from someone, through the speaker system, fire alarms, etc. I stayed on
around the general vicinity where I worked. This is the main reason why
I stayed longer on the floor." Interview 1000576 (NIST 2004)
Similarly, 16 percent (n=59) of WTC 2 evacuees on September 11, 2001 also evacuated in 1993. In
WTC 2, however, only 75 percent (n=42) felt that they made the right decision in 1993 (compared to
95 percent in WTC 1), possibly due to the fact that many more waited to evacuate in 1993 in WTC 2
(69 percent (n=39)) than did so in WTC 1. Only 31 percent (n=17) who reported their decision evacuated
immediately from WTC 2 in 1993, keeping in mind that the bomb had a more significant impact on
WTC 1 in 1993.
Nearly every respondent who compared the 1993 bombing evacuation to the 2001 evacuation indicated
that the 1993 evacuation was slower, more difficult (presence of smoke in the stairwells and floors), and
more stressful. Having participated in the 1 993 evacuation, those occupants generally felt much better
about the progress and conditions during the 2001 evacuation. As a consequence of their 1993
experience, however, several 2001 survivors who had direct 1993 experience reported having diminished
confidence in building announcements (or the lack thereof), exemplified by this occupant from a floor in
the 40s in WTC 2:
"I ... was there during the 1 993 bombing. I did evacuate —with a group
of people who had no clue as to where we were going. What I learned
from that experience was to not trust the Port Authority's
announcements. Because of experiencing the '93 bombing, I felt a strong
conviction to get out this time. Had I not experienced that, I might have
listened to the Port Authority announcements and stayed put." Interview
1000048 (NIST 2004)
146 NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
10.3.6 Occupant Activities
Occupants of WTC 1 and WTC 2 performed a number of activities before and during evacuation on
September 11, 2001. These activities included gathering personal belongings, milling with other
occupants, seeking additional information, and calling family or friends, among others. In general, these
activities either delayed the start of their evacuation (pre-evacuation activities) or interrupted their
evacuation once it was in progress and are described below.
Pre-Evacuation Activities
Occupants performed a variety of activities prior to entering the stairwell (or elevator in WTC 2) to begin
their evacuation. An activity such as gathering occupants or warning them to leave has a variable amount
of time associated with it, depending upon the floor space to search and how reluctant others are to leave,
whereas an activity such as gathering belongings usually requires under a minute to complete. Other
activities noted by occupants were milling, seeking information (such as looking out the window, making
phone calls, searching the internet), helping others, following emergency procedures, fire fighting, and
working or closing up a work desk. The delay reported by survivors in starting their evacuation was
predominately one to five minutes, while a small number delayed for more than an hour. The
distribution of evacuation initiation delay times was discussed in detail in Chapter 6.
The majority of the survivors, 70 percent of those in WTC 1 and 75 percent in WTC 2, engaged in the
activity of talking to others, or milling. The behavior of milling is used by occupants to discuss what they
witnessed and to arrive at conclusions for what happened and what to do next.
For instance, an occupant in the 90s of WTC 2 looked for other
coworkers after he heard a sound coming from WTC 1. "I moved back
out of my office where I could get in sight with other people to share my
observations, to discuss what ought to be done, and to hear what the
consensus would be." Interview 1000689 (NIST 2004)
After hearing an explosion and feeling the building rock, an occupant in
the 60s of WTC 1 turned to a coworker. "I discussed with my coworker
what we were going to do. The noise of the explosion, the shaking of the
building, and seeing the paper falling was what made me decide it was
something serious. We left our conference room and went into the
hallway to get to the stairs." Interview 1000053 (NIST 2004)
Near the floors of impact in WTC 1 , a few survivors worked to secure an area of refuge or perform fire
fighting activities. The floors near the impact area were much more difficult to navigate due to the smoke
and damage. Because of this, some occupants remained in an office, surrounding the openings with wet
towels, before being rescued by Port Authority personnel.
In both towers, the face-to-face data captures the fact that some occupants delayed their evacuation start
to wait to receive evacuation instructions, as part of their emergency procedures. Fourteen out of
124 respondents in WTC 1 and 12 out of 69 respondents in WTC 2 waited for instructions. The
percentages are higher for WTC 2 because it included those occupants who heard the public address
system announcement while still on their floor before they began evacuation.
NIST NCSTAR 1-7, WTC Investigation 147
Chapter 10
On at least one floor in the 30s in WTC 1, occupants followed the
emergency procedures. "We knew that we were supposed to go out to
the hallway and that there would be someone there with information or
information would be given via loudspeaker. We went out to the
hallway to listen for any instructions." Interview 100009 (NIST 2004)
Despite the number of pre-evacuation activities of some occupants, most occupants in WTC 2 chose to
begin evacuating before their tower was attacked. Eighty-six percent to 91 percent 63 of the survivors in
WTC 2 began their evacuation before their building was hit at 9:02:59 a.m. Similar to the telephone
interview results (which are generalizable), 12 percent of the face-to-face respondents in WTC 2 stayed
until their tower was attacked. Of those who stayed (8 interviews out of 66 from WTC 2), five held
positions of corporate leadership or fire warden. The remaining three made multiple decisions to leave
before the plane hit, but were delayed by announcements, phones ringing, and the desire to gather
belongings.
Helping
September 11, 2001 showed that people are willing to help out others during an emergency, even if they
do not know the person ahead of time. Those occupants in need of help during the evacuation included
occupants who use canes or wheelchairs, overweight, elderly, pregnant women, and people with asthma.
Many of those who supported occupants in need throughout their evacuation were strangers. These
helpers often remained with the occupant in need throughout their entire evacuation, even though they
were putting themselves at risk. Short-term examples of helping behaviors mentioned in both towers
were comforting coworkers in the stairwell, leading occupants in need to the elevators, helping an
occupant carry belongings, and encouraging others to keep evacuating in times of stress or exhaustion.
Occupants also exhibited major feats of heroism on September 11, 2001 by caring for an individual
throughout their entire evacuation, sometimes assisting strangers down more than 60 stories to the
building.
Examples of this heroism include:
A group of coworkers helped to rescue a wheelchair user from a floor in
the 60s in WTC 1 . 15 minutes after the impact of their tower, WTC 1 ,
they located the evacuation chair and transferred the occupant into it.
"We began our descent sliding the evacuation chair, step by step, [with]
two helpers holding the chair on top and two helpers holding the chair on
the bottom. [Soon] we began carrying [the occupant] down the
staircase." Interview 1000123 (NIST 2004)
Two occupants assisted an overweight occupant in WTC 1 in navigating
the stairwell from a floor in the 50s. Towards the bottom of the stair, the
occupant's "legs would give out and [the occupant] would fall in our
arms. Our descent slowed dramatically as we practically had to carry
[the occupant] down each flight of stairs." Interview 1000093 (NIST
2004)
63
The discrepancy reflects a small difference among responses to two independent questions: 'How long did it take you to
evacuate?' and 'Did you evacuate WTC 2 prior to WTC 2 being attacked?', as discussed previously.
148 NIST NCSTAR 1 - 7, WTC Investigation
Discussion and Analysis
On a floor in the 80s of WTC 1 : "I heard the secretary, who was in
flames, calling for help near the door. I extinguished the flames with my
bare hands, along with a colleague who was using a sweater." Interview
1000113 (NIST2004)
However, an offer of help was not always accepted. One survivor from a floor in the 60s in WTC 1
recalled:
"There were some people that stopped in descending because they
needed to catch their breath. We stopped about three times and offered
help, which was not accepted." Interview 1000878 (NIST 2004)
Resting During Evacuation
Some occupants felt the need to rest at certain points during their evacuation. An occupant coming down
from the 90s by the stairs was forced to walk roughly one-half mile before eventually leaving the
building. 64 Occupants often chose to rest inside the stairway, either on the steps or, more frequently, on a
stairwell landing. Resting during evacuation was noted by respondents beginning evacuation from either
the high or medium strata of each building. Most often, resting was reported by occupants with
respiratory problems, obesity, or other physical or medical conditions.
From the face-to-face interviews, respondents' experience of resting on the stairs is captured in the
following quotes:
An occupant in WTC 1 had to travel from a floor in the 70s. Around
floor 25, the occupant's "legs were really hurting and it was difficult to
walk. I slowed down and stopped about three times. The first time, I let
people get ahead of me. I stopped at a landing for approximately two
minutes. I proceeded down but then stopped for the second time,
somewhere between the 15th and tenth floors. I waited there for three or
four minutes to think what was the best way to continue. There was
water gushing from the fifth floor onto the staircase. I stopped and began
to think about the safest way to get down and decided to proceed more
slowly and take a firmer grip on the rail." Interview 1000111 (NIST
2004)
In WTC 2, a group of coworkers found themselves also having to rest
multiple times during their evacuation from a floor in the 90s. "On the
78th floor, we felt tired. My coworker, who is a diabetic, hadn't eaten
breakfast yet. We stopped on the stairs to rest for a minute. A lot of
people seemed to be stopping, probably due to the heat and needing to
catch their breath. On the 64th floor, another coworker turned to me and
said [he or she] was getting tired and didn't feel well. I said 'We will
take a little break' and I gave [the occupant] mints." Interview 1000526
(NIST 2004) The group stopped at least two additional times because a
64
At roughly 30 degree slope, an occupant descending stairs from 1,000 vertical ft also travels over 1,500 horizontal ft, for a total
combined distance of over 2,500 ft, or approximately one-half mile. Travel from stair exit and transfer floors increases this
distance.
NIST NCSTAR 1-7, WTC Investigation 149
Chapter 10
coworker wasn't feeling well and didn't think he/she could make it down
the stairs.
Occupants also observed other occupants in the stairs who were resting inside the stairwell. Face-to-face
interview respondents sometimes mentioned specific categories of individuals who were resting in the
stair, including people having trouble breathing (asthma), overweight, elderly, helpers, and firefighters.
Leaving the Stairs
In WTC 1 and WTC 2, occupants left the stairs for a variety of reasons. While occupants left the stairs
onto floors throughout the buildings, the most commonly reported floor was one of the skylobbies (floor
44 or 78). In addition, there were a number of occupants who left the stairs at the lower floors of WTC 1,
which may be attributed to either debris from the collapse of WTC 2 or the fire department rest station for
occupants somewhere between floors 12 and 20.
In WTC 1, the most frequent reasons cited for leaving the stairs were an instruction to do so from
firefighters, Port Authority, or building security (33 percent) and the stair condition (41 percent),
including crowding or smoke/dust/jet fuel in the stairs. Before the collapse of WTC 2, occupants were
directed out of certain stairs to either switch to another stair immediately or wait on the floor for a certain
period of time.
One occupant who originated on a low floor in WTC 1 stated that, "We
were directed by a man (I am unsure if he was a security guard). He was
telling people to go to the 18th floor in order to scatter the traffic from
the stairwell. When we arrived at the 18th floor, we got out of the
stairwell and into a vacant floor space." This occupant later stated that,
"He said he was instructed to pile up the people into the vacant floors to
control the flow of traffic." Interview 1000769 (NIST 2004) The
occupant demanded to know why this was happening and decided to find
another stair to take out of the building. A group then followed this
occupant to another stairwell.
After the collapse of WTC 2, the occupants near the lower part of WTC 1 were faced with an onslaught of
debris from the collapse. For this reason, occupants were again instructed to leave the stairs and if not
instructed, took it upon themselves to switch stairs at times.
Two occupants helping an overweight colleague down Stair A in WTC 1
after the collapse of WTC 2 "were told by firemen and rescue workers
that [they] had to go up to the fourth floor [because] the exit was
blocked." Interview 1000093 (NIST 2004)
Another frequent reason for leaving the stairs in WTC 1 , which may have been unknowingly caused by
either instruction or stair condition, was occupants following other occupants (9 percent). These people
felt comfortable following others out of the stair, without always knowing the reason for leaving or who
had initiated the stair move. Other reasons for leaving the stairs in WTC 1 include a jammed exit at the
transfer floor (76) of Stair A, occupant's being uncomfortable, the presence of a mobility impairment,
retrieving something, helping another person, and seeking information.
1 50 NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
In WTC 2, the most frequent reasons for leaving the stairs were instructions to leave the stairs by Port
Authority or building security (10 percent) personnel, the public address system announcement made
before WTC 2 impact (32 percent), and elevator usage (13 percent). Some occupants in WTC 2 were
being instructed out of the stairwells at the skylobbies minutes before the announcement was made.
When the announcement was made to occupants throughout the building, some occupants inside the stair
walked out onto a floor to hear the announcement more clearly and/or to react to the information given by
the announcement. Also, some occupants decided to take the elevators to a different point in the building
(to either evacuate or return to their offices) both before and after the announcement was made. Other
reasons for leaving the stairwell in WTC 2 included following the crowd, occupants being uncomfortable,
and seeking information.
10.3.7
Aids and Constraints to Evacuation
Evacuation incentives are interpersonal interactions or physical features of the built environment that
helped people to evacuate. Analysis of the telephone interview data reveals that evacuees that among
occupants who received help from other people, 9 percent reported that they were assisted by floor
wardens, 44 percent reported that they were helped by police or firefighters, and 65 percent reported that
they were assisted by "others."
For occupants who were helped by building features, 33 percent of survivors in WTC 1 and 17 percent of
those in WTC 2 reported that they were helped by photoluminescent markings. The discrepancy between
towers may be due to the fact that lights were lost in WTC 1 after WTC 2 collapsed, thereby
demonstrating the usefulness of the photo luminscent qualities. Additionally, occupants who used
elevators in WTC 2 would not have observed the photoluminescent paint.
Conversely, certain conditions presented constraints to evacuation. Table 10-1 shows the most frequently
reported constraints to evacuation from the telephone interview data. Three areas were reported by more
than half of the evacuees in WTC 1: crowded stairwells, emergency responders in the stairwells, and
injured or disabled persons in the stairwells. The findings from the causal model for normalized stairwell
evacuation time in WTC 1, however, provide a scientific basis for refuting the occupant's perception that
firefighters entering the building adversely affected the overall flow down the stairwells. While more
than half (63 percent) reported that the firefighters / police in the stairwells were a constraint, an occupant
who encountered firefighter or police did not have a significantly slower stairwell travel time than an
occupant who did not encounter firefighter or police in the stairwell, all other factors being held constant.
Table 10-1. Constraints to evacuation.
WTC 1
WTC 2
Stairwells were too crowded
73 % (321)
69 % (206)
Firefighters/police in stairwell
63 % (275)
27 % (80)
Injured/disabled in stairwell
52 % (226)
33 % (99)
Lack of direction/information
24% (104)
29% (106)
Locked doors
16% (72)
7 % (25)
Poor lighting
1 1 % (48)
4% (15)
Bad/missing signage
5 % (23)
5% (18)
Source: NIST WTC telephone survey data.
NISTNCSTAR 1-7, WTC Investigation
151
Chapter 10
A possible explanation for this seeming inconsistency is that while an occupant was required to stop and
stand to the side in the stairwell as firefighters and police moved past, they were able to temporarily
increase their speed to catch back up to where they would have been before they had met the firefighters
or police. This explanation is contingent upon the occupants descending the stairwell at less the
maximum speed, which was found to be the case earlier in this chapter (the travel speed was about one-
half what would be expected in a non-emergency evacuation).
Elevator Usage in WTC 2
At least 1 8 percent of the survivors from WTC 2 reported using the elevators for at least part of their
evacuation, including those who used elevators from the basement levels. Those who used elevators in
WTC 2 did so for various reasons. While most occupants used elevators for evacuation, some had
decided to find a quick way to return to their office once the 9:00 a.m. announcement was given. Elevator
usage was not necessarily dominated by people with mobility impairments, but used by all people with
the intent to evacuate the building quickly. However, one occupant using the elevators out of need started
out in the 90s of WTC 2.
"I wanted to get out of the building as quickly as possible. I was taking a
new medication and knew I should not walk down the stairs." Interview
1000553 (NIST 2004) This occupant took an elevator from the 95th
floor to the 78th and then switched to another elevator at the 78th floor to
travel to the lobby level.
Despite the availability of elevators for occupants with mobility impairments in WTC 2, however,
approximately the same percentage of mobility-impaired occupants chose to use elevators in WTC 2
(19 percent), when compared to the surviving population overall (16 percent).
Face-to-face interview respondents also refer to elevator usage after the plane hit. An occupant, injured
on a floor in the 70s, was evacuated via elevator by a firefighter and a security guard, along with two
other injured occupants, after WTC 2 has hit by the airplane.
"As we were walking down, we saw a fireman coming up and told us to
get to 40 and that someone would take us in an elevator down to the
lobby." Interview 1000562 (NIST 2004)
The use of elevators in WTC 2 saved many lives due to the fact that occupants from floors 78 - 107 in
WTC 2 used both stairwells and elevators in order to move below the impending impact region prior to
the WTC 2 attack. In order to estimate the total number of occupants able to descend below the impact
zone prior to the WTC 2 attack, the following assumptions were made: (a), no occupant began evacuating
WTC 2 prior to 9:02:59 a.m.; (b), no elevators were usable; and (c), the evacuation rate of WTC 2
mirrored the observed evacuation rate of WTC 1 in Figure 10-2 (starting at 8:46:30 a.m.). Under these
three assumptions, over 3,000 people would have remained in the building as it collapsed, with over
2,000 occupants remaining trapped above the 78th floor. Therefore, self evacuation (starting to evacuate
prior to 9:02:59 a.m.) and the use of elevators during that time period is estimated to have saved roughly
3,000 lives in WTC 2 on September 1 1, 2001.
1 52 NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
Footwear
Evacuation speed on the stairs can be significantly affected by the choice of footwear worn by the
occupants, especially women. High-heeled shoes, especially heels higher than 3 in., dramatically
decrease movement capability on the stairs (Templer 1992). People, mostly women, removed their shoes
in the stairwell in both towers. This presented a potential hazard for other occupants in the stair, who had
to maneuver around the pile of shoes, as well as for the occupants without shoes walking through the
damaged portions of the building.
In WTC 1, an occupant noticed a pile of shoes in a stair near the 28th
floor. "There was a pile of shoes that accumulated from people kicking
them off. Some of the people around me were tripping on them and
warning others to watch out for them." Interview 1000042 (NIST 2004)
Firefighters (Interviews 1000081 and 1000540 (NIST 2004)) and Port Authority personnel (Interview
1000071 (NIST 2004)) occasionally instructed occupants to put their shoes back on.
Transfer Hallways
As described in Chapter 2.2, the stairwells did not descend in a straight vertical alignment in WTC 1 and
WTC 2. The horizontal connections, more numerous and lengthy in Stairwells A and C than Stairwell B,
could extend more than 100 ft and require several 90 degree turns. In addition to slightly increasing
evacuation time (compared to a design with no horizontal transfers), the transfer hallways introduced
uncertainty in the minds of the evacuees regarding the correctness of their evacuation path.
An occupant from a floor in the 60s in WTC1 described the transfer
hallways: "As I descended the stairs down to the Mezzanine Level, once
or twice I had to exit the staircase through a door and go down a corridor
in order to reconnect to the same stairwell. I found this to be extremely
disconcerting. Everyone who did this, stopped before they exited the
staircase to make sure they were doing the right thing. This slowed us
down and there was concern that the door would lock behind us."
Interview 1000053 (NIST 2004)
10.3.8 Emergency Responders and Building Authorities
Emergency Responders
In addition to organizing the response to the attacks on the two towers and assisting occupants during
their evacuation, building staff and emergency responders had to use the stairwells to attempt access to
impacted floors in the buildings. This resulted in small groups of firefighters in bulky bunker gear moving
against the flow of occupants down the stairs. This phenomena is often referred to as counterflow. Police
and fire department involvement with building occupants was identified by survivors as both an
evacuation aid and constraint. The police and firefighters were identified as an evacuation aid by
44 percent of the occupants in WTC 1 and 30 percent of the occupants in WTC 2 and as a constraint to
evacuation by 62 percent of the occupants in WTC 1 and 27 percent of the occupants in WTC 2. The
lower numbers should be expected since before WTC 2 was hit, emergency responders naturally
NIST NCSTAR 1-7, WTC Investigation 1 53
Chapter 10
concentrated their efforts in WTC 1 . Many of those who evacuated early from WTC 2 never encountered
police or firefighters in the building.
Inside WTC 1 stairwells, firefighters were sighted by interviewees in all three stairwells, with a
concentration on Stair B (from the face-to-face interviews). Also, from the face-to-face interviews,
firefighters were sighted as high as the 60s in WTC 1. Interviews 1000576 and 1000645 (NIST 2004) For
those meeting firefighters in the stairs, some mentioned slowing down, crowding, and even stopping
several times. An occupant in stairwell A of WTC 1 was both slowed down and reassured by the
firefighters.
"We encountered firemen ascending, starting at about the 35th floor. I
came to a dead stop numerous times in Stairwell A for about 5 minutes
each time to aid the firemen to get up to the problem. The firemen were
easy-going, and attempted to put people at-ease. They were also
extremely winded and sweating profusely from their climb." Interview
1000103 (NIST 2004)
Figure 1 0-7 shows an FDNY firefighter ascending a 44 in. ( 1 . 1 m) stairwell in WTC 1 on
September 11, 2001. Figure 10-8 shows how an occupant and a firefighter in bunker (turnout) gear may
not be able to pass one other in a 1.1 m (44 in.) stairwell without either the occupant or the firefighter
moving somewhat to the side. Figure 10-8 was not taken in a stairwell from the WTC complex, but was
intended to be a generic demonstration of the counterflow phenomena.
Figure 10-7. Firefighter and occupants using a 44 in. stairwell in WTC 1 on
September 11, 2001.
154
NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
The firefighters provided a sense of
security to some occupants in the towers.
In some cases, occupants went so far as to
show their appreciation. An occupant
from WTC 1 Stairwell A recalled
encountering firefighters near the 20th
floor.
"Firemen were going up and they were
looking exhausted. People were
cheering the firemen and some of the
firemen said, 'Don't stop, you're
almost there.' As we moved aside, we
handed the firemen water." Interview
1000722 (NIST2004)
Throughout the event, emergency
responders supported the occupants by
performing a variety of actions, including
directing occupants to change stairs
(especially in WTC 1 when exits were
blocked by debris from the collapse of
WTC 2), providing directions on how to
exit through the Concourse, giving out
snacks, water, and oxygen (from air
bottles) to occupants in need.
■* *MI
^d
w
A fH
II
-
Source: NIST.
Figure 10-8. Occupant and firefighter in bunker
gear passing in a generic 44 in. stairwell.
Firefighters reportedly established a rest
station somewhere between floors 12 and
20 (Interview 1000543 [NIST 2004]). Firefighters here instructed occupants to drop certain occupants off
at this floor for assistance, as well as suggesting that evacuees stop on this floor if they required rest or
assistance.
In addition to walking up toward the fire floors, many firefighters moved injured occupants to safety. An
occupant inside Stairwell C of WTC 1 heard instruction from floor above to keep to the right.
The respondent then "noticed that the firemen in uniforms were bringing at least two people" down the
stairs. "One gentleman had his arms severely burned and one lady who had her head halfway burned and
was screaming with pain. After the fireman came down, others came up - several of them. They had
gear on their back and each one had an extra hose on their shoulder. I was on the right and was letting
them up, and a particular firemen patted me on the left shoulder and said, 'Don't worry, you will be ok.'"
Interview 1000697 (NIST 2004)
NISTNCSTAR 1-7, WTC Investigation
155
Chapter 10
Impact of Authorities on Occupant Evacuation
In both towers, occupants followed directions from those in a position of authority. In an office
emergency, higher rank can mean multiple things. On the floors, occupants followed instructions to leave
from supervisors or fire wardens, even though supervisors and wardens may have had the same
information as the employee about what was going on. Also, occupants followed directions throughout
the building from the building and fire officials who were familiar with the building layout, such as the
Port Authority employees, police, firefighters, and building security. Analysis of face-to-face interview
data revealed that these instructions included when and where to evacuate the stairs, when and where to
change stairs, when to move to the right to let firefighters go up and injured down, or where to travel upon
leaving the stairwell.
Workplace Authority
In an emergency, there is usually a combination of reasons why a person begins their evacuation
process. 5 However, there is usually one significant reason that finally makes them decide to leave or that
weighs more heavily on their decision than the others. When occupants in both towers were asked to
name the one thing that made them decide to evacuate, 14 percent of the occupants in WTC 1 and
21 percent of the occupants in WTC 2 said that the reason they evacuated was being told to evacuate.
(Telephone Interviews, NIST 2004)
At the first awareness that something was wrong, the occupants could only rely on each other (or
themselves, if alone) to understand what had happened and to decide what to do next. Fire alarms
sounded in certain areas of WTC 1 (and possibly WTC 2); however, occupants did not generally report
the fire alarms as their reason to evacuate. The others on their floor were their subordinates, coworkers,
or superiors (supervisors or fire wardens). From a total of 208 face-to-face interviews, 86 people
mentioned being instructed to evacuate, as well as their reaction to that instruction. Most respondents
began their evacuation when told to leave, whether by a superior or co-worker. Although the percentage
was higher when direction came from a superior, it is difficult to draw any conclusion from the face-to-
face data since it is not statistically -based and cannot be generalized to the entire population of the
buildings. Only one face-to-face interview respondent reported telling their superior to evacuate, and in
that isolated case, the superior ignored the instruction.
An occupant from the 40th floor in WTC 1 left the floor due to persuasion by the fire warden:
"As I was typing the email message, I heard a loud voice say 'Leave.' I
kept typing the message when a fire warden grabbed my arm and pulled
me out of the seat." The fire warden also pointed this occupant in the
direction of the elevators and stairs in addition to instructing the
occupant to leave the floor. Interview 1000802 (NIST 2004)
After hearing an explosion from WTC 1 , an occupant with workplace authority on a floor in the 60s of
WTC 2 instructed others on the floor to leave:
Nelson and MacLennan. "Emergency Movement." In The SFPE Handbook of Fire Protection Engineering, 2 n Edition.
NFPA, Quincy, MA. 1995.
156 NIST NCSTAR 1 - 7, WTC Investigation
Discussion and Analysis
"I saw fire and burning debris and smelled jet fuel. I ran out of my office
and yelled to staff and co-workers to order them to evacuate. We ran to
the stairwell." Interview 1000556 (NIST 2004)
And, even after WTC 2 was hit, occupants on a floor in the 50s of WTC 2 remained on their floor
discussing what had just happened and what they should do next:
"Our facilities manager told me it was time to go. We did a quick check
of the floor to make sure no one was left behind. In the hallways we saw
that everyone was heading in one direction and we followed them to the
stair." Interview 1000557 (NIST 2004)
It would appear from the face-to-face interviews that occupants were likely to follow the instruction given
by their superior in an emergency. However, even if the final decision to evacuate was prompted by the
instruction to evacuate, other factors may have been involved in making that decision, including seeing
the plane strike into WTC 1, past experience, or other observations of the event.
Building Authority
Similar to the FDNY role in the WTC towers on September 11, 2001, the building authority played an
important role in providing guidance to occupants about where to go once they left their floors, which the
occupants frequently followed. The Port Authority personnel were observed giving instructions and
directions to occupants at the skylobbies, Mezzanine, and Concourse areas of the WTC towers. Their
instructions for some occupants consisted of when and where to leave the stairs, whether to use the
elevators, and in WTC 2, when to return to their offices. Port Authority direction consisted of how and
where to go through the building and Concourse area in order to leave the WTC complex. In most cases,
occupants followed the instructions and welcomed the directions given by the building authority in the
towers.
Many noted the tremendous help that building personnel provided at the base of the building. An
occupant in WTC 2 noted that the building authorities were present throughout the Concourse:
"Trade center people were directing us into the Concourse because you
couldn't go out Liberty street - they had all those exits closed because
there was debris flying all over the street. Security guards were like a
human chain telling us which direction to go. We followed the security
guards' direction . . . towards Borders." Interview 1000842 (NIST 2004)
An occupant in WTC 1 noted the Port Authority giving directions to occupants on the Mezzanine:
"As soon as I arrived onto the Mezzanine floor, I saw a chief operating
officer giving directions to get onto the stairs [escalator] and go down.
He also said not to look out onto the plaza because it was unsafe and
dangerous." Interview 1000639 (NIST 2004)
In addition to directing occupants throughout the buildings, there were two cases from the face-to-face
interviews where occupants were rescued from their floors by the help of the building authority in
NIST NCSTAR 1-7, WTC Investigation 1 57
Chapter 10
WTC 1. In both cases, the occupants were faced with worsening conditions on their floor and were
guided to the stairwells by the official.
From a floor in the 80s in WTC 1 : After preparing their office with wet
towels around the doors to prevent the heavy smoke from seeping in, the
respondent "saw a beam of light through the smoke, which turned out to
be a fireman and a building worker. They told us to drop everything and
follow them into the hall that led to the staircase." Interview 1000055
(NIST 2004)
After two unsuccessful attempts to open the jammed door from a floor in the 80s into stairwell A, the
heavy smoke on the floor drove the group of coworkers back to their office to wait, until:
"We heard one of our coworkers in the hall [say that] the Port Authority
fellow had opened up the doors to Stairwell A for us." Interview
1000137 (NIST 2004)
10.3.9 Occupant Experience
Experience from the 1993 bombing and other past evacuation experience affected some occupants'
actions in 2001. For example, many of those involved in both 1993 and 2001 recalled long evacuation
times in 1993, including occupant congestion, smoke in the stairs, and not being allowed to return to the
building for weeks after the event. Because of this experience, some occupants started their evacuation as
soon as WTC 1 was hit and noted that their experience in 1993 was the principal reason. On the other
hand, some occupants performed specific activities that they wished they had performed in 1993, such as
calling home so their parents/family wouldn't worry about them and taking certain belongings in case
they couldn't return to the building right away.
Even though occupants of the towers may not have been present for the 1993 bombing, the bombing
event may have played a role in their 2001 evacuation. Many of those present in 1993 shared their
experiences with other employees both before and during the 2001 attack:
"I was not in the building in 1993, but a lot of people who had been there
during the 1993 bombing were very helpful because they were exiting
faster. They knew that every second counted based on that previous
experience and I took my cue from them." Interview 1000518 (NIST
2004)
"I wasn't at the building in 1993, but I knew that it took over an hour to
get down the stairs, which influenced my decision to use the elevators."
Interview 1 00073 1 (NIST 2004)
10.3.10 Mobility-Impaired Occupants
When evacuating a high-rise building, many different physical and medical conditions can affect travel
ability on stairs. As the total distance traveled to reach an exit increases, the number of people unable to
successfully complete the evacuation without resting or requiring assistance increases. In the WTC, stair
travel challenged occupants with certain conditions, such as wheelchair use, pregnancy, asthma, visual
158 NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
impairment, physical impairment, obesity, arthritis, and old age. The current NYC Building Code defines
a physical disability as any one of the following:
• Impairment requiring the use of a wheelchair; or
• Impairment causing difficulty or insecurity in walking or climbing stairs or requiring the use of
braces, crutches, or other artificial supports; or
• Impairment caused by amputation, arthritis, spastic condition or pulmonary, cardiac, or other ills
rendering the individual semi-ambulatory; or
• Total or partial impairment of hearing or sight causing insecurity or likelihood of exposure to
danger in public places; or
• Impairment due to conditions of aging.
Evacuation of WTC 1 and WTC 2 during the 1 993 bombing identified an inefficiency at the World Trade
Center in evacuating the mobility-impaired. The Report of the World Trade Center Review Committee in
1995 indicated: "Evacuation of persons with disability from the World Trade Center was slow and
arduous, the [sic] Committee proposes that methods of evacuation should be studied to provide equally
safe egress for these building occupants." (New York City 1995)
Despite introduction of evacuation chairs and a buddy system for pre-identified mobility-impaired
occupants, 51 percent of the occupants in WTC 1 and 33 percent of the occupants in WTC 2 in 2001,
noted that injured and disabled persons in the stairwell were a constraint to evacuation. However,
occupants were quick to aid these individuals by guiding them throughout their evacuation or simply
moving to the side of the stairwell to let those who were injured and others in need pass by when they
could.
In WTC 1, "someone was being carried down in some kind of
handicapped apparatus and was strapped in. The occupant was being
carried by two fellow occupants around the 30s. We stopped to allow
them access." Interview 1000834 (NIST 2004)
In some cases, occupants noted passing slower mobility-impaired individuals in the stairs and even
slowing or stopping behind them.
In WTC 2, "we saw an [occupant] who was hyperventilating. [The
occupant] was walking down the stairs with assistance. We slowed
down and came to a stop [because] we couldn't get around the two
[occupants]." Interview 1000556 (NIST 2004)
In WTC 1, two occupants were helping an overweight occupant evacuate
the building. During one of the helper's descent down the stairs, the
helper noted that "we took up the entire width of the stairway and no one
could get around us until we came to a landing." Interview 1000093
(NIST 2004)
NIST NCSTAR 1-7, WTC Investigation 1 59
Chapter 10
Finally, some occupants reported mobility-impaired occupants waiting on the stairs and/or landings for
others to help them or to be rescued by the fire department. Many respondents recalled two specific
occupants from WTC 1 who weren't able to evacuate the building in time. These included an occupant in
a wheelchair waiting with a friend on the stairs and an occupant with arthritis waiting at a fire department
rest stop somewhere between floors 12 and 20.
Mobility-Impaired Occupants and Mortality below the 92nd floor in WTC 1
Several occupants perished assisting mobility-impaired colleagues (Fahy and Proulx 2003) and many
more occupants and emergency responders risked their lives assisting mobility-impaired colleagues who
successfully evacuated. However, for occupants where a likely mechanism contributing to unsuccessful
evacuation could be identified, being trapped by debris on the starting floor, delayed evacuation initiation
(of statistical outlier magnitude), or performing emergency response building responsibilities accounts for
the majority of the below the impact region deaths.
10.4 EVACUATION SIMULATIONS
The purpose of modeling evacuation from the World Trade Center towers was to obtain evacuation times
for a variety of scenarios in order to provide additional context with which to understand the
September 11, 2001 evacuation of WTC 1 and WTC 2. Table 10-2 shows each of the six general egress
simulations performed, along with details regarding the evacuation type (full building evacuation or
phased evacuation, also known as defend-in-place), the number of evacuees included in the simulation,
the input response delay, which models were used, and any other relevant information about the
simulation. There were five full building evacuations and one phased evacuation simulated. The full
building evacuation simulations explored the effect on the total evacuation time (or number of successful
evacuees if the time was fixed) with respect to the number of simulated evacuees, the presence or absence
of building damage (observed on September 11, 2001), and the type of model used for the simulations. It
should be noted that none of the models used in this analysis have been validated for emergency
evacuation in buildings as large as 1 10 stories.
10.4.1 Egress Simulation Results
Phased Evacuation
The following section is a summary of the egress simulation results. For a more complete discussion of
egress modeling inputs, assumptions, limitations, and results, refer to Appendix D: Egress Modeling. The
purpose of simulation 1 , phased evacuation, was to understand not only the time necessary to perform a
phased evacuation, but to compare the results using three different egress models: Simulex, EXIT89, and
buildingEXODUS.
Table 10-3 shows the total phased evacuation times for each model. The three models simulate a total
phased evacuation time between 3!/2 and 4 minutes for all 600 occupants, assuming that evacuees start
evacuating immediately. If the evacuees were randomly assigned a delay time between and 10 minutes,
the simulated total evacuation time was between 1 1 and 1 1 14 min for all 600 occupants. Phased
evacuation is an efficient strategy to quickly remove occupants most at risk quickly from 'routine'
emergencies, or those that fit within the design envelope of the life safety systems.
160 NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
Table 10-2. Egress simulation matrix.
Scenario
Full Building or
Phased
Evacuation
Number of
Evacuees
Simulated
Response
Delay
Model(s)
Used
Notes
1
Phased
600
Zero
- 10 min
Simulex
EXODUS
EXIT89
• Occupants travel three floors
below fire floor
2
Full Building
19,800
Zero
0- 10 min
EXODUS
EXIT89
• Fully-occupied, without visitors
• No damage
3
Full Building
25,000
Zero
- 10 min
EXODUS
EXIT89
• Fully-occupied, with visitors
• No damage
4
Full Building
8,800
Zero
- 10 min
EXODUS
EXIT89
• September 11, 2001 population
• No damage
5a
Calibration
Simulation
7,200
6-30 min
EXODUS
• Stop and Go
• Ground - Floor 90
• Damage above Floor 90
5b
FullWTC 1,
with damage
16,000
6-30 min
EXODUS
• Stop and Go
• Ground - Floor 90
• Damage above Floor 90
6a
Calibration
Simulation
7,400
2-17 min
(Above
Floor 77)
6-30 min
(Ground to
Floor 76)
EXODUS
• Stop and Go
• Ground - Floor 76
• No Damage
6b
Elevator, WTC 2
19,800
Zero
ELVAC
• 14 Minute Elevator Simulation
6c
Full WTC 2,
With Damage
17,260
2-17 min
(Above
Floor 77)
6-30 min
(Ground to
Floor 76)
EXODUS
• Stop and Go
• Ground to Floor 107 for first 16
min
• Ground to Floor 77 for time range
1 7 min - 72 min
Ta
Die 10-3. Results for phased evacuation simulations.
Occupant Type,
Evacuation Time (s)
Evacuation Model
Characteristics
No Delay
0-10 min Delay
Simulex
All office staff
60 % men
40% women
240
690
EXODUS
5 % males, age 17-29
38% males, age 30 -50
21 % males, age 51 -80
3 % females, age 17-29
22 % females, age 30 - 50
11 % females, age 51 - 80
243
660
EXIT89
All medium body size
Emergency speed
210
690
NISTNCSTAR 1-7, WTC Investigation
161
Chapter 10
Full Building Evacuation
The purpose of simulations 2, 3, and 4 was to explore the effect of additional building occupants on the
total building evacuation time. The three populations selected were (1) the NIST estimate of building
population on September 11, 2001, (2) the NIST estimate of a fully-occupied tower without visitors
(19,800 occupants), and (3) the NIST estimate of a fully-occupied tower including visitors
(25,500 occupants). Note that in 2005, Port Authority estimated that the maximum population of WTC 1
or WTC 2 would not likely have exceeded 20, 000. 66 Figure 10-9 shows the model results for the three
building populations using the results from the buildingEXODUS model. The slope of the regression was
identical whether comparing the September 11, 2001 population evacuation time to fully-occupied
building evacuation time ([112 min - 55 min] / (19,800 - 8,800) = 5.2 min per thousand additional
occupants) or compared to fully-occupied with visitors building evacuation time ([142 min - 55 min] /
(25,600 - 8,800) = 5.2 min per thousand additional occupants). Thus, for a given building geometry, and
a number of other important simulation assumptions, adding additional occupants to a building population
resulted in a linear increase in total building evacuation time.
Extrapolation of Evacuation Time for a Fully-Occupied Tower on September 11, 2001
In Chapter 4.1 of this report, NIST estimated that the number of successful evacuees from WTC 1 on
September 11, 2001 was 7,500, the total of which required roughly 100 minutes to exit the building. As
shown in Figure 10-9 and Table 10-4, the buildingEXODUS model estimated that a fully-occupied
building (population 25,500) required approximately 2.6 times as long (142 / 55 = 2.6) as a building with
a September 11, 2001 occupant load; therefore, on September 11, 2001, a fully-occupied WTC tower with
visitors, may have required roughly (100 x 2.6) = 260 min (over 4 hours) to fully evacuate.
Using the same logical approach, if WTC 1 had been occupied by approximately 20,000 occupants on
September 11, 2001 (fully-occupied without visitors), the evacuation would have taken (1 12 min / 55 min
= 2.0 x 100 min = 200 min) over 3 hours to complete.
Table 10-4. Total building evacuation time (simulated) for various occupant loads.
Evacuation Model
Evacuation
Initiation Delay
Input
Evacuation Time
(min): 8,800
occupants
Evacuation time
(min): 19,800
occupants
Evacuation
time (min):
25,500
occupants
EXODUS
10 Minute Delay
55
112
142
No Delay
52
110
141
EXIT89
10 Minute Delay
71-74
92-113
119-139
No Delay
58-78
97-117 67
114-140
66 Bhol, Saroj. PANYNJ (September 21, 2005). Email from S. Bhol to S. Sunder in response to NIST question.
The underlying theory for people movement in EXIT89 is based upon the work of Predtechenskii and Milinskii. They
observed an inverse relationship between density and speed for three different types of movement: emergency, normal, and
comfortable. The effect of crowding (density) on overall evacuation speed was greater when no delay time was assumed. A
delay time when 19,800 occupants were present, according to EXIT89, spaces the occupants out and increases overall
evacuation efficiency. This was not found to be the case for all occupant loads (25,500).
162
NIST NCSTAR 1-7, WTC Investigation
Discussion and Analysis
Full Building Evacuation of a WTC Tower
160
140
"c 120
'E
I* 100
80
c
o
3
o
ill 40
60
20
A
25,585 - full capacity
with visitors
19,800 - full capacity
8,800 - 9/11 capacity
III
5000 10000 15000 20000
Number of occupants in Tower
25000
30000
Figure 10-9. Simulation of full building evacuation of a WTC tower with different
occupant loads.
Estimated Occupant Mortality from a Fully-Occupied WTC Tower on September 11, 2001
Scenarios 5 and 6 were simulated in order to estimate the consequences of a fully-occupied, without
visitors (total building population: 19,800) from WTC 1 or WTC 2 on September 11. Each tower was
simulated with a September 1 1 population, and model assumptions and inputs (such as evacuation
initiation delay time) were refined in order to roughly match the actual outcome on September 11. After
the buildingEXODUS model results were determined to resemble gross characteristics of the
September 1 1 evacuation, more occupants (fully-occupied building, without visitors) were added, and the
results are presented below.
WTC 1 was assumed to have no passable stairwell above floor 91 starting at time zero. Elevators were
assumed to be rendered inoperable. The simulation was ended at 1 03 minutes and any simulated
occupant remaining in WTC 1 was assumed to have perished. The simulation shows that 69 percent of all
occupants (13,600 occupants out of 19,800) would have evacuated WTC 1 prior to collapse: percent
(0 out of 3,800) from above the impact zone and 85 percent (13,600 out of 16,000) from below the impact
zone. Table 10-5 summarizes the results of scenarios 5 and 6.
NISTNCSTAR 1-7, WTC Investigation
163
Chapter 10
Tabl
e 10-5. Simulated evacuation results for fully-
September 11, 2001.
occupied WTC 1 and WTC 2 on
Building
Total
Number of
Occupants
at t = 0.0
Potential
Number
of
Evacuees
Occupants
Remaining
in Building
at Collapse
Occupants
Trapped
Above
Floors of
Impact
Percentage of Occupants Who
Successfully Evacuated Relative to
Where They Started
Total
Below
Impact
Above
Impact
WTC 1
19,800
16,000
6,200
3,800
69%
85%
0%
WTC 2
19,800
17,260
8,377
3,900
58%
75%
44%
Total
39,600
33,260
14,577
7,700
63%
Occupants of WTC 2 were assumed to use elevators for a period of 1 6 minutes, after which all elevators
were assumed inoperable. After 16 minutes, when WTC 2 was attacked, it was assumed that no
occupants survived above floor 77. The simulation was ended after 72 minutes, when WTC 2 collapsed.
The simulation showed that 1 1,423 of the 19,800 occupants, or 58 percent of the initial population, would
have successfully evacuated. Of the 1 1,423 simulated occupants, 8,883 evacuated using the stairwells,
while 2,540 simulated occupants evacuated using elevators. Of the 8,377 occupants who remained in the
building at the time of collapse, 3,900 simulated occupants were trapped above floor 77. Another 4,477
were evacuating from below the 78th floor when WTC 2 collapsed: 1,231 simulated occupants were
initially above the 77th floor, while 3,246 did not evacuate despite starting on a floor below the
78th floor. The initial population of occupants below the 78th floor was 12,783. Therefore, (3,246 /
12,783) 25 percent of the occupants who started below the impact floors did not successfully evacuate.
Similarly, 3,900 of the initial 7,017, or 56 percent of the occupants at or above the impact region at
8:46:30 a.m. in WTC 2 did not successfully evacuate. The simulation showed that 549 occupants initially
above floor 77 successfully evacuated using stairwells, while only 3 of the 549 simulated occupants
originated above floor 100.
When combining the evacuation modeling results for WTC 1 and WTC 2, assuming 19,800 occupants in
each tower and assuming that the aircraft impacts and collapses occurred at the same times as they did on
September 11, 2001, about 14,000 occupants may have perished, not including any emergency
responders, aircraft passengers, or bystanders.
10.4.2 Egress Modeling Conclusions
The egress modeling revealed three principal insights into the evacuation of WTC 1 and WTC 2 on
September 11, 2001. First, while not an appropriate strategy for an emergency the scale of
September 11, 2001, phased evacuation, under certain circumstances, moves occupants most at risk to a
place of relative safety much more quickly and with less total impact upon building tenants than full
building evacuation. Second, additional occupant load in a WTC tower (compared to September 11,
2001) resulted in a linear increase in total building evacuation time. Third, assuming that the ratio of
observed to simulated evacuation time of 2.6 would extend to the fully-occupied with visitors
(25,500 occupants), full building evacuation scenario, WTC 1 would have taken over four hours to fully
evacuate under the conditions of September 11, 2001.
164
NIST NCSTAR 1-7, WTC Investigation
Chapter 11
Findings
The evacuations of World Trade Center (WTC) 1 and WTC 2 on September 11, 2001, were documented
and analyzed. In order to provide context for the 2001 evacuations, relevant historical egress events were
explored. The evolution of the egress and communication systems, as well as the emergency procedures
within WTC 1 and WTC 2 was documented. Over 1 ,000 interviews, using advanced interrogatory
methods, were conducted. Hundreds of published accounts from a variety of sources were collected and
analyzed. Emergency call records, emergency communication transcripts, Occupational Safety and
Health Administration complaints, and Port Authority of New York and New Jersey design records and
plans were analyzed. Analysis and compilation resulted in the following conclusions.
11.1 BUILDING POPULATION AND DEMOGRAPHICS
• There were 8,900 ± 750 people in WTC 1 at 8:46:30 a.m. on September 11, 2001. Of those,
7,470 (or 84 percent) survived the attacks, while 1,462 - 1,533 occupants were killed in WTC 1.
The National Institute of Standards and Technology (NIST) Investigation found that at least 107
occupants were killed below the 92nd floor.
•
•
There were approximately 8,540 ± 920 people in WTC 2 at 8:46:30 a.m. on September 11, 2001.
Of those, 7,940 (or 93 percent) survived the attacks, while 630 - 701 occupants were killed in
WTC 2. The NIST Investigation found that at least 1 1 occupants were killed below the 78th
floor.
Sixty-seven percent of WTC 1 occupants and 51 percent of WTC 2 occupants had started
working at the World Trade Center during the four years prior to September 11, 2001.
Two-thirds of WTC 1 and WTC 2 occupants participated in at least one fire drill in the twelve
months prior to September 11, 2001. Seventeen percent of the occupants did not participate in a
fire drill during that time period, and 17 percent did not remember whether they had participated
in a fire drill during that time period.
• Nearly all occupants (93 percent) who participated in a fire drill were instructed about the
location of the nearest stairwell.
11.2 EVACUATION
• Approximately 87 percent of WTC occupants, and over 99 percent of those below the floors of
impact, were able to successfully evacuate.
- At 9:02:59 a.m., when WTC 2 was hit and 17 minutes after WTC 1 was hit, 21 percent of
survivors had exited WTC 1 and 41 percent of survivors had exited WTC 2.
NIST NCSTAR 1-7, WTC Investigation 1 65
Chapter 11
- By 9:37 a.m., 22 minutes before collapse, 95 percent of survivors had exited WTC 2.
- At 9:58:59 a.m., when WTC 2 collapsed, 88 percent of survivors had exited WTC 1.
- By 10:12 a.m., 16 minutes before collapse, 95 percent of survivors had exited WTC 1.
• Occupants of WTC 1 overwhelmingly initiated their own evacuation. The NIST Investigation
found no evidence that public address system announcements were heard by occupants of the
building, although the fire command station was attempting to make announcements.
• Self-evacuation and use of elevators for 16 minutes in WTC 2 saved roughly 3,000 lives.
• During the last 20 minutes before each building collapsed, the evacuation rate in both buildings
had slowed to about one-fifth the immediately prior evacuation rate. This suggests that for those
seeking and able to reach and use the undamaged exits and stairways, the egress capacity (the
number and width of exits and stairways) was adequate to accommodate survivors.
• In WTC 1, the average surviving occupant spent approximately 48 seconds per floor in the
stairwell. That does not include any time prior to entering the stairwell, which was often
substantial.
• Some occupants of WTC 1 delayed or interrupted their evacuation resulting in over 100 deaths
below the impact region.
• The NIST Investigation found no evidence to indicate that anyone who was above the 91st floor
in WTC 1 after the airplane impact survived. This was due to the fact that the stairwells and
elevators were destroyed and helicopter rescue, despite several attempts by both occupants and
aircraft, was not possible.
• In WTC 2, approximately 75 percent of the occupants above the 78th floor at 8:46:30 a.m. had
successfully descended below the 78th floor prior to the airplane impact at 9:02:59 a.m. This
occurred in spite of conflicting announcements, first urging people to return to their offices
around 9:00 a.m., and then informing them that they may initiate an evacuation if conditions
warranted around 9:02 a.m.
•
•
The NIST Investigation found evidence that 18 people successfully used Stairwell A in WTC 2 to
leave the building after being on or above the 78th floor when Flight 175 hit the building. One of
the 18 later died from injuries sustained on September 11, 2001. Additionally, at least two people
went to or above the 78th floor after having been below the 78th floor at the time of impact in
order to assist trapped people and did not survive the collapse of WTC 2.
Minutes prior to the collapse of WTC 2, an NYPD Emergency Services Unit (ESU) officer
radioed from a floor in the 20s to the outside that he was having trouble ascending the stairwell
due to the large number of occupants descending (Interview 24 NYPD [NIST 2004]). While the
origin of the occupants remains unknown, only 1 1 occupants who started evacuating below the
impact region were known not to have survived.
1 66 NIST NCSTAR 1-7, WTC Investigation
Findings
•
Computer egress modeling indicated that a full capacity evacuation of a single WTC tower with
approximately 20,000 occupants required a minimum evacuation time of 1 h and 50 min. Given
that the actual evacuation time on September 11, 2001, was about 100 min without elevator use, a
full capacity evacuation of the WTC towers by 20,000 people would likely have required
somewhat greater than 3 h (2 times 100 min). To achieve a significantly faster total evacuation at
full capacity would have required increases in egress capacity (number and width of exits and
stairways).
Egress modeling indicates that, had WTC 1 and WTC 2 been fully-occupied on September 11,
2001, with approximately 20,000 occupants each, about 14,000 occupants may have perished in
the building collapses.
11.3 DELAYS IN EVACUATION
•
•
Occupants in WTC 1 delayed starting their evacuation because environmental cues (information
from the physical environment that something was terribly wrong) and floor (increased distance
to safety) caused people to seek additional information. Next, the act of seeking additional
information, that is "milling" about to make sense out of the situation, led people to take actions
to prepare to evacuate. Finally, taking those actions to prepare to evacuate delayed the initiation
of actually evacuating. Although there were other factors that had lesser influence on what
people did, the paths of causal influence that defined the main process that led people to delay in
the evacuation of WTC 2 on September 11, 2001 were identical to those for WTC 1, with one
decided difference. Perceived risk was predicted by environmental cues and initial floor and also
contributed to seeking additional information and taking pre-evacuation actions in WTC 2, while
the effect of perceived risk was substantially lower in WTC 1.
• Starting floor (increased distance to safety) substantially increased the odds that people would
encounter environmental cues. Floor also increased delay in starting evacuation (this relationship
is elaborated upon in much greater detail in Chapter 10 of this report), which, in turn, also
increased the chances that people would encounter environmental cues. Observation of
environmental cues (information from the physical environment that something was terribly
wrong) in the stairwell had a large and direct effect on increasing the amount of time that people
spent, on average, in their evacuation stairwell. In addition to this multi-step process with
environmental cues as the key predicting variable, interrupting the process of evacuation for any
reason increased the amount of time, on average, that people used to descend their evacuation
stairwell.
• Contrary to the perceptions of the occupants, counterflow in WTC 1 was determined by causal
modeling analysis not to be a significant predictor of increased total evacuation time while in the
stairwells when compared to other factors, including evacuation interruption and environmental
cues.
•
Occupants in WTC 1 delayed starting their evacuation because environmental cues (information
from the physical environment that something was terribly wrong) and floor (increased distance
to safety) caused people out to find additional information, most likely information about what
was going on and what they should do about. Next, the act of seeking additional information, that
NIST NCSTAR 1-7, WTC Investigation 1 67
Chapter 11
is "milling" about to make sense out of the situation, led people to take actions to prepare to
evacuate. Finally, taking those actions to prepare to evacuate delayed the initiation of actually
evacuating. Although there were other factors that had lesser influence on what people did, the
main paths of causal influence that defined the main process that led people to delay in the
evacuation of WTC 2 on September 1 1th were identical to WTC 1 with one decided difference.
This was that perceived risk was predicted by environmental cues and initial floor and also
contributed to seeking additional information and taking pre-evacuation actions in WTC 2 while
the effect of perceived risk was substantially lower in WTC 1. This was likely the case because
WTC 1 was hit without warning, and only the people in WTC 2 had time to wonder (perceive) if
their tower was going to be a target.
11.4 EMERGENCY MANAGEMENT AND PLANNING
• Building occupants, 9-1-1 operators, fire department dispatch, WTC building officials, and Port
Authority personnel lacked adequate situational awareness despite nearly constant cross-
communications. Many opportunities to communicate important information in a timely manner
were missed, such as telling building occupants the general location of the impact region or
whether to evacuate or not.
• Faced with an uncertain situation, occupants of both buildings received conflicting feedback /
advice from a variety of sources (including 9-1-1 operators, FDNY, family and friends, and the
Port Authority) regarding whether to evacuate, whether to break windows, and what the nature of
their situation was.
• World Trade Center occupants were inadequately prepared to encounter horizontal transfers
during the evacuation process and were occasionally delayed by confusion as to whether the
hallway led to a stairwell and confusion about whether the transfer hallway doors would open or
be locked.
In addition to an announcement in WTC 2 just prior to the airplane impact, announcements were
made from the fire command station in the lobby of WTC 2 after the aircraft impact, although the
NIST Investigation found no evidence that any surviving occupants heard these announcements.
The decision to establish the primary evacuation route underground through the Concourse (mall)
and out up to street level by WTC 5 (commonly recalled as being by the Borders Bookstore)
prevented a significant number of injuries and/or deaths.
The first "first responders" were colleagues and regular building occupants. Acts of everyday
heroism saved many people whom traditional emergency responders would have been unable to
reach in time.
11.5 OCCUPANTS WITH MOBILITY IMPAIRMENTS
•
•
•
Approximately 1,000 surviving occupants (a projection of the 6 percent reported in the telephone
interviews) of WTC 1 and WTC 2 had a limitation that impacted their ability to evacuate,
including recent surgery or injury, obesity, heart condition, asthma, elderly or otherwise requiring
1 68 NIST NCSTAR 1-7, WTC Investigation
•
Findings
assistance to walk, pregnancy, and others. The most frequently reported disabilities were recent
injuries and chronic illnesses; the number of occupants requiring use of a wheelchair was very
small, relative to the frequency of other mobility impairments.
While many mobility-impaired individuals were able to successfully evacuate, often with
assistance from co-workers or emergency responders, others were temporarily removed from the
stairwells in order to allow more able occupants to evacuate the building (such as the rest station
low in WTC 1 [somewhere between floors 12 and 20]). It remains unclear whether all of the
mobility-impaired occupants and the helpers were able to successfully evacuate on September 11,
2001. No evidence of a similar rest station in WTC 2 was found.
While the mobility status of every decedent known to be below the impact region (107) in WTC 1
could not be determined, it does not appear that mobility-impaired individuals were significantly
overrepresented among the decedent population.
Mobility-impaired occupants were not universally accounted for by existing evacuation
procedures, as some were left by colleagues (later assisted by strangers), while others chose not to
identify their mobility impairment to any colleagues.
NIST NCSTAR 1-7, WTC Investigation 1 69
Chapter 11
This page intentionally left blank.
1 70 NIST NCSTAR 1-7, WTC Investigation
Chapter 12
References
1912. The Equitable Building Fire. Quarterly of the National Fire Protection Association 5, no. 4:408-
445.
1977. Triangle Shirtwaist Toll Of 145 Is Still Largest for U.S. Industrial Fires. Fire Engineering 130, no.
8:55-58.
Adler, J. 2002. Five Who Survived. Newsweek, September 9.
ASME 2000. Safety Code for Elevators and Escalators ASME 17.1, American Society of Mechanical
Engineers, New York.
Blalock, H. 1972. Social Statistics. New York: McGraw Hill.
BOMA 2001. Experience Exchange Report, Building Owners and Managers Association, New York.
Bouchard, J. 1982. NFPA Summary Report on the Holiday Inn Fire, Kerney, Nebraska. NFPA Journal
76, no. 6:20-24.
Breasted, Mary. 1977. 100,000 Leave New York Offices As Bomb Threats Disrupt City; Blast Kills One
and Hurt Seven. New York Times.
Bukowski, R.W. and Kuligowski, E.D. 2003. The Basis for Egress Provision in U.S. Building Codes,
Proceedings of InterFlam 2004, Interscience Communications, London.
Bukowski, R.W. and Moore, W. D. 2003. Fire Alarm Signaling Systems. 3rd (Third) Edition. National
Fire Protection Association, Quincy, MA.
Bukowski, R. W. and W. D. Moore. 2003. Fire Alarm Signaling Systems. Quincy, MA: National Fire
Protection Association.
Cochrane, W. G. 1977. Sampling Techniques. New York: John Wiley & Sons.
Couchon, D. 2001. For Many on Sept. 11, Survival Was No Accident . USA Today, December 19.
Demers, D. 1980. Familiar Problems Cause 10 Deaths in Hotel Fire. NFPA Journal 74, no. 1:52-56.
DiMarco, D. 2004. Tower Stories, The Autobiography of September 11, 2001, Revolution Publishing,
New York.
Fahy, R. F. 2001. Verifying the Predictive Capability of EXIT89. Paper presented at 2nd International
Symposium on Human Behaviour in Fire.
NIST NCSTAR 1-7, WTC Investigation 171
Chapter 12
Fahy, R. F. 1999. User's Manual, EXIT89 v 1.01, An Evacuation Model for High-Rise Buildings.
Quincy, Ma, National Fire Protection Association.
Fahy, R. F. and G. Proulx. 2003. Database of First-Person Accounts from Survivors of the World Trade
Center Evacuation on September 11, 2001 conducted as part of report NIST NCSTAR 7-A.
Fasullo, E. 1995. The Port Authority of New York and New Jersey, New York, personal communication
to R. Visconti, New York City Dept. of Buildings. January 27. WTCI-159-P
Fisher, R. P., K. H. Brennan, and M. R. McCauley. 2002. The Cognitive Interview Method to Enhance
Eyewitness Recall. In Memory and Suggestibility in the Forensic Interview, edited by Eisen, M. L., J.
A. Quas, and G. S. Goodman (Mahweh, New Jersey: Lawrence Erlbaum).
Fisher, R. P., K. L. Falkner, M. Trevisan, and M. R. McCauley. 2000. Adapting the cognitive interview to
enhance long-term (35 years) recall of physical activities. Journal of Applied Psychology 85, no.
2:180-189.
Fruin. J. J. 1987. Pedestrian Planning and Design, Revised Edition. Elevator World, Incorporated,
Mobile, AL.
Galea, E. R., P. Lawrence, S. J. Blake, S. Gwynne, and Westeng H 2004. A Preliminary Investigation of
the Evacuation of the WTC North Tower Using Computer Simulation. 167-180. London, England,
Interscience Communications Limited. 3rd International Symposium on Human Behaviour in Fire.
Geiselman, R. E., R. P. Fisher, D. Mackinnon, and H. Holland. 1986. Enhancement of eyewitness
memory with the cognitive interview. American Journal of Psychology 99:385-401.
Gwynne, S., E.R. Galea, P. J. Lawrence, M. Owen, and L. Filippidis. 1999. A Review of the
Methodologies used in the Computer Simulation of Evacuation from the Built Environment, Building
and Environment, 34, 741-749.
Gwynne, S., E. R. Galea, P. J. Lawrence, and L. Filippidis. 2001. Modeling Occupant Interaction with
Fire Conditions Using the buildingEXODUS Model. Fire Safety Journal, 36, 327-357.
Gwynne, S., E. R. Galea, P. J. Lawrence, M. Owen, and L. Filippidis. 1998. A Systematic Comparison of
Model Prediction Produced by the buildingEXODUS Model and the Tsukuba Pavilion Evacuation
Data, Applied Fire Science, Vol. 7, No. 3, 235-266.
Hayne, W. G. 1945. Report of the Collision of U.S. Army Air Force Bomber with the Empire State
Building at 78th and 79th Floors. 1-15. New York, New York Board of Fire Underwriters.
IES 2000. Simulex Technical Reference; Evacuation Modeling Software. Integrated Environmental
Solutions, Inc.
IES 2001. Simulex User Manual; Evacuation Modeling Software. Integrated Environmental Solutions,
Inc.
Isner, M. S. 1990. Five Die in High-Rise Office Building Fire. NFPA Journal 84, no. 4:50-59.
1 72 NIST NCSTAR 1 - 7, WTC Investigation
References
Isner, M. S. and T. J. Klem. World Trade Center Explosion and Fire, New York, New York, February 26,
1993. Fire Investigation Report. 1993a. Quincy, MA, National Fire Protection Association.
Isner, M. S. and T. J. Klem. 1993b. Explosion and Fire Disrupt World Trade Center. NFPA Journal:9l-
104.
Ivins, Molly. 1977. Most Stay Calm As Evacuations Halt Workday. New York Times.
Jones, W. W., G. P. Forney, R. D. Peacock, and P. A. Reneke. Technical Reference for CFAST: An
Engineering Tool for Estimating Fire and Smoke Transport. Tech. Note 1431. 2000. Natl. Inst.
Stand. Technol.
Keating, J. and E. Loftus. Post Fire Interviews. 1984. Development and Field Validation of the
Behavioral Sequence Interview Technique. NBSIR 84-477. 1984. Nat. Bur. Stand. (U.S.).
Klote, J. H. 1993. A Method of Calculation of Elevator Evacuation Time. Journal of Fire Protection
Engineering 5, no. 3:83-95.
Levy, M. P. 1965. The Port of New York Authority, New York, personal communication to M.
Yamasaki, Minora Yamasaki & Associates, New York. September 29.
Memorandum from John M. Kyle (PONYA) to Guy F. Tozzoli (PONY A), May 9, 1966. "World Trade
Center - New York City Building Code - Proposed Revision."
Murphy, Dean E. 2002. September 11: An Oral History. New York: Doubleday.
New York 1973. Fire Safety in Office Buildings, New York City - Local Law 5, The City of New York.
January 18.
New York 1984. Fire Safety in Buildings, Including Hotels, New York City - Local Law 16, The City of
New York. .
New York 1995. The Report of the World Trade Center Review Committee. The City of New York.
Nillson, L. 2000. Remembering actions and words. In The Oxford handbook of Memory, edited by
Tulving, E. and F. I. M. Craik (New York: Oxford University Press).
NIST 2004a. NIST WTC Face-to-Face Interview Data Set. 2004. Gaithersburg, MD, Natl. Inst. Stand.
Technol.
NIST 2004b. NIST WTC Emergency Responder Interview Data Set. 2004. Gaithersburg, MD, Natl. Inst.
Stand. Technol.
NYCBC 1968. Building Code - Local Law No. 76 of the City of New York, New York.
NYCBC 2001. Building Code of the City of New York, 2001 Edition, Gould Publications, Binghamton,
NY.
NIST NCSTAR 1-7, WTC Investigation 1 73
Chapter 12
PANYNJ Undated. "WTC - CHANGES TO EVACUATION SYSTEMS AFTER 1993." The Port
Authority of New York and New Jersey. WTC-1 15-P
PANYNJ 1995. The Port Authority of NY & NJ World Trade Center Fire Safety Plan (Towers One and
Two), The Port Authority of New York and New Jersey, New York. WTCI-13-NYC.
PANYNJ 1996. WTC Fire Safety 1996. Video. WTCI-463-P
PANYNJ 2001a. Port Authority of New York and New Jersey Radio Communications Transcripts from
September 11,2001.
PANYNJ 2001b World Trade Center Emergency Procedures Manual 2001 - Confidential. The Port
Authority of New York and New Jersey.
PANYNJ 2001c. World Trade Center Space Book Plan, Building 1 and Building 2, Port Authority Real
Estate Department, The Port Authority of New York and New Jersey, New York.
Powers, W. R. 1975. One World Trade Center Fire New York, N.Y., February 13, 1975. 1-15. Boston,
MA, National Fire Protection Association.
Powers, W. R. 1970. One New York Plaza Fire, New York, N.Y., August 5,1970. 1-15.. New York, New
York Board of Fire Underwriters.
Predtechenskii, V. M. and A. I. Milinskii. 1978. Planning for Foot Traffic in Buildings. New Delhi:
Amerind Publishing Co. Pvt. Ltd.
Schank, R. C. and R. Abelson. 1977. Scripts, Plans, Goals, and Understanding. Nillsdale, NJ: Erlbaum.
Sharry, J. A. 1974. South America Burning. Fire Journal, 1974, 23-33.
Solomon, J., 1968. Emory Roth and Sons, New York, personal communication to M. Levy, The Port of
New York Authority, New York. January 25. WTCI-477-P
Solomon, J., 1975. Emory Roth and Sons, New York, personal communication to M. Levy, The Port of
New York Authority, New York. February 18.
Templer, J. 1992. Staircase: History and Theories. MIT Press , Cambridge, MA.
Willey, A. E. 1972. High Rise Building Fire, Sao Paulo, Brazil. Fire Journal, 1972, 7-108.
Williams, F. D. 1990. SLAM: The Influence of S.L.A. Marwill on the United States Army. In TRADOC,
(Ft. Monroe, VA: U. S. Army).
1 74 NIST NCSTAR 1-7, WTC Investigation
Appendix A
Telephone Interview Script
I would like to start by getting some background information. What year did you first start working at the
World Trade Center? RANGE: 1975 - 2001
$E 1975 2001
DK9998
RF 9999
«YRWRK»
On September 1 1th, 2001, were you in any of the following positions with the World Trade Center?
Port Authority Staff 1
Fire Safety Staff 2
Floor Warden or Searcher 3
Maintenance or Security Staff 4
NONE OF THESE X
DK8
RF 9
«ROLES_01»
«ROLES_02»
«ROLES_03»
«ROLES_04»
During the year from September 1 1th, 2000 to September 1 1th, 2001, how many fire drills did you take
part in at the World Trade Center?
$E0 99
NONE 00 => SWLOC
DK98 => SWLOC
RF 99 => SWLOC
«FIRED»
During these drills, were you ever instructed about the location of the emergency stairwell nearest to your
office?
YES 1
NIST NCSTAR 1-7, WTC Investigation 1 75
Appendix A
NO 2 => SWLOC
DK 8 => SWLOC
RF 9 => SWLOC
«DEXIT»
How many emergency stairwells were you shown?
ONE 1
TWO 2
THREE 3 => LVFSW
OTHER, SPECIFY 7 O
DK8
RF 9
«HMEXT»
«0 HMEXT»
Before September 1 1th, had you learned in other ways about the locations of the three emergency
stairwells?
YES 1
NO 2
DK 8
RF 9
«SWLOC»
SKIP IF NO FIRE DRILLS
=> USESW
Else=>+1
ifFIRED=00,98-99
«SOUTl»
During any of the fire drills, did you leave your floor using one of the stairwells?
YES 1
NO 2 => USESW
DK 8 => USESW
RF 9 => USESW
«LVFSW»
1 76 NIST NCSTAR 1 - 7, WTC Investigation
Telephone Interview Script
Which stairwells did you use?
STAIRWELL A 1
STAIRWELL B 2
STAIRWELL C 3
OTHER, SPECIFY 7 O
DK8
RF 9
«WHSW1_01»
«WHSW1_02»
«WHSW1_03»
«WHSW1_04»
«0_WHSW1»
Which side of the building was the stairwell located on?
=>+l
if NOT WHS W 1=8
NORTH 1
SOUTH 2
EAST 3
WEST 4
OTHER, SPECIFY 7 O
DK8
RF 9
«WHSL1»
«0_WHSL1»
Had you ever used any of the emergency stairwells prior to September 1 1th?
=> DHELP
ifLVFSW=l
YES 1
NO 2 => DHELP
DK 8 => DHELP
RF 9 => DHELP
«USESW»
NIST NCSTAR 1-7, WTC Investigation 1 77
Appendix A
SKIP FOR NO DRILLS AND NO USE OF STAIRWELLS
=> AEVOF
Else=>+1
if FIRED=00,98,99 AND USESW>1
«SOUT2»
Which stairwell did you use?
STAIRWELL A 1
STAIRWELL B 2
STAIRWELL C 3
OTHER, SPECIFY 7 O
DK8
RF 9
«WHSW2_01»
«WHSW2_02»
«WHSW2_03»
«WHSW2_04»
«0_WHSW2»
SKIP IF NO FIRE DRILLS
=> AEVOF
Else=>+1
ifFIRED=00,98-99
«SOUT3»
When you were evacuating on September 1 1th, how helpful was your experience during these drills?
=>+l
ifFIRED=00
Very Helpful 1
Somewhat Helpful 2
Somewhat Unhelpful 3
Very Unhelpful 4
DK8
RF 9
178 NIST NCSTAR 1 - 7, WTC Investigation
Telephone Interview Script
«DHELP»
Prior to September 1 1th, were you aware of the evacuation procedures for your floor?
YES 1
NO 2 => FLWAR
DK 8 => FLWAR
RF 9 => FLWAR
«AEVOF»
Prior to September 1 1th, what was the evacuation procedure you were told to follow?
LEAVE BUILDING IMMEDIATELY 1
GO TO ELEVATOR LOBBY 2
GO TO FLOORS UP OR DOWN 3
GO TO ROOF 4
STAY WHERE YOU ARE 5
OTHER, SPECIFY 7 O
DK8
RF 9
«EVACP»
«0_EVACP»
Did you know that there was a Floor Warden for your floor?
=>+l
ifROLES=l-4
YES 1
NO 2
DK 8
RF 9
«FLWAR»
The next questions ask about 3 different time periods. The first series of questions asks about when you
first became aware that something had happened at the World Trade Center. This is a period of just a few
seconds. The next series of questions asks about the time from when you first became aware that
something had happened, to the time you first entered a stairwell or elevator to exit the building. The
third series of questions asks about what happened during your evacuation, meaning the time from when
NIST NCSTAR 1-7, WTC Investigation 1 79
Appendix A
you first entered a stairwell or elevator until you exited the tower. At the end of the interview, I will ask
you if there is anything else about your experience on September 1 1th that you would like to contribute.
CONTINUE 1 D
«IFAWA»
Now thinking back to the morning of September 1 1th, how did you first become aware that something
had happened at the World Trade Center?
$E 1 9
HEARD SOMETHING (BOOM, CRASH, EXPLOSION, BLAST, ROAR, RUMBLING, ALARM)
01
SAW SMOKE OR FLAMES 02
SAW DEAD BODIES 03
SAW A PLANE 04
SAW DEBRIS 05
FELT SOMETHING (BUILDING MOVING, IMPACT, SHAKING, SWAYING, ROCKING, JOLT,
EARTHQUAKE) 06
FELL DOWN/FELL OFF CHAIR 07
WARNED BY SOMEONE AROUND ME 08
CONTACTED VIA PHONE 09
CONTACTED VIA EMAIL 1
PUBLIC ADDRESS SYSTEM 11
NEWS MEDIA (TELEVISION, RADIO) 12
OFFICE FURNITURE OR FIXTURES FALLING 1 3
FURNITURE OR OTHER ITEMS FALLING OVER/DOWN 1 4
OTHER, SPECIFY 97 O
DK98
RF 99
«FAWAR»
«0_FAWAR»
What were you doing when you first became aware that something had happened to the World Trade
Center? PROBE: Anything else?
$E 1 9
WORKING INDEPENDENTLY 01
IN MEETING 02
ON PHONE 03
180 NIST NCSTAR 1-7, WTC Investigation
Telephone Interview Script
CHECKING/WRITING EMAIL
WAITING FOR ELEVATOR 05
RIDING IN ELEVATOR 06
CHATTING WITH COWORKERS
EATING/HAVING COFFEE 08
ENTERING BUILDING 09
OTHER, SPECIFY 97 O
DK98 X
RF 99 X
«ACTV1_01»
«ACTV1_02»
«ACTV1_03»
«ACTV1_04»
«ACTV1_05»
«ACTV1_06»
«ACTV1_07»
«ACTV1_08»
«ACTV1_09»
«ACTV1_10»
«0 ACTV1»
04
07
At the moment when you first became aware that something had happened at the World Trade Center, did
you notice any of the following? FOLLOW UP: Was that in your immediate area or outside the Tower?
Did Not Notice
Noticed in Immediate Area
Noticed Outside the
Tower
Smoke
Fire or Flames
Fireballs
Collapsed walls
Jet Fuel
Severely or fatally injured people
Sprinklers going on
A fire alarm sounding
Power outage or flickering lights
Fallen ceiling tiles
Extreme heat
NISTNCSTAR 1-7, WTC Investigation
181
Appendix A
«NOT01_01»
«NOT01_02»
TIME PERIOD: 1
Were there any disaster related events going on around you at this time?
=> WHTW2
ifOR[NOT01-NOTll]=2-3
YES 1
NO 2 => WHTW2
DK 8 => WHTW2
RF 9 => WHTW2
«OEVEN»
TIME PERIOD: 1
What was going on?
ENTER RESPONSE 1 O
DK8
RF 9
«GOING»
«0_GOING»
TIME PERIOD: 1
Were you still in<WHTOW>at this time? IF YES, SELECT APPROPRIATE CHOICE IF NO, ASK
WHICH TOWER THEY WERE IN
TOWER 1 1
TOWER 2 2
DK8
RF 9
«WHTW2»
TIME PERIOD: 1
And were you still on the<WHFLO>floor at this time? RANGE: 1st - 1 10th FLOOR IF YES,
SELECT/ENTER FLOOR IF NO, ASK WHICH FLOOR THEY WERE ON AND SELECT/ENTER IT
$E 1 110
BASEMENT 990
1 82 NIST NCSTAR 1 - 7, WTC Investigation
Telephone Interview Script
CONCOURSE/LOBBY 99 1
PLAZA992
IN ELEVATOR 993
OTHER, SPECIFY 997 O
DK998
RF 999
«WHFL2»
«0_WHFL2»
TIME PERIOD: 1
At the moment when you first became aware that something had happened to the World Trade Center,
approximately how many people were with you? RANGE: - 999 PEOPLE WE WANT THE
NUMBER OF PEOPLE THAT WERE IN THE SAME LOCATION AS THE RESPONDENT. (IN
THEIR LINE OF SIGHT)
$E 999
NONE 00 =>YOUIN
DK 98 => YOUIN
RF 99 => YOUIN
«PEOPl»
TIME PERIOD: 1
Were any of these people injured at that time as a result of the event?
YES 1
NO 2
DK 8
RF 9
«PEOESr»
TIME PERIOD: 1
Were you injured at that time, as a result of the event?
YES 1
NO 2 => ORISK
DK 8 => ORISK
RF 9 => ORISK
«YOUIN»
NIST NCSTAR 1-7, WTC Investigation 1 83
Appendix A
TIME PERIOD: 1
Would you say your injury was a . . .
An injury that did not impact your ability to evacuate, 1
An injury that did impact your ability to evacuate but was not life threatening, or 2
A life threatening injury 3
OTHER, SPECIFY 7 O
DK8
RF 9
«NATIN»
«0_NATIN»
TIME PERIOD: 1
Still thinking about the moment when you first became aware that something had happened at the World
Trade Center, did you believe that other people were in danger of being killed?
YES 1
NO 2
DK 8
RF 9
«ORISK»
TIME PERIOD: 1
Did you believe you were in danger of being killed?
YES 1
NO 2
DK 8
RF 9
«YRISK»
TIME PERIOD: 2
Now please think about the time period between when you first became aware that something had
happened and when you first entered a stairwell or elevator to leave the tower. During this entire time
period, were you given any additional information about what was going on? AFTER BECOMING
AWARE OF THE EVENT, BUT BEFORE EVACUATION
YES 1
NO 2 => SEEKI
DK 8 => SEEKI
1 84 NIST NCSTAR 1 - 7, WTC Investigation
Telephone Interview Script
RF 9 => SEEKI
«GETIN»
TIME PERIOD: 2
Who gave you this information? PROBE: Anyone else?
MANAGER/SUPERVISOR 1
COWORKER INSIDE BUILDING 2
FAMILY/FRIEND OUTSIDE BUILDING 3
POLICE/FIREFIGHTER 4
FLOOR WARDEN 5
MEDIA PERSON (TV/RADIO) 6
OTHER, SPECIFY 7 O
DK8 X
RF 9 X
«WHINF_01»
«WHINF_02»
«WHINF_03»
«WHINF_04»
«WHINF_05»
«WHINF_06»
«WHINF_07»
«0_WHINF»
TIME PERIOD: 2
What information did you get? PROBE: Any other information?
INFORMATION ABOUT WHAT HAD HAPPENED 1
INSTRUCTIONS TO LEAVE 2
INSTRUCTIONS TO STAY 3
OTHER, SPECIFY 7 O
DK8 X
RF 9 X
«WHATI_01»
«WHATI_02»
«WHATI_03»
«WHATI_04»
NIST NCSTAR 1-7, WTC Investigation 1 85
Appendix A
«0_WHATI»
TIME PERIOD: 2
How did you get this information? PROBE: Any other way?
FACE TO FACE 1
TELEPHONE 2
EMAIL/BLACKBERRY 3
PA ANNOUNCMENT 4
TV/RADIO 5
OTHER, SPECIFY 7 O
DK8 X
RF 9 X
«HOWGT_01»
«HOWGT_02»
«HOWGT_03»
«HOWGT_04»
«HOWGT_05»
«HOWGT_06»
«0_HOWGT»
TIME PERIOD: 2
And during this same time period, did you try to get additional information about what was going on?
AFTER BECOMING AWARE OF THE EVENT, BUT BEFORE EVACUATION
YES 1
NO 2 => ORIS2
TRIED, BUT WAS UNABLE TO GET INFORMATION 3 => ORIS2
DK 8 => ORIS2
RF 9 => ORIS2
«SEEKI»
TIME PERIOD: 2
Who did you go to for this information? PROBE: Anyone else?
MANAGER/SUPERVISOR 1
COWORKER INSIDE BUILDING 2
FAMILY/FRIEND OUTSIDE BUILDING 3
186 NIST NCSTAR 1 - 7, WTC Investigation
Telephone Interview Script
POLICE/FIREFIGHTER 4
FLOOR WARDEN 5
MEDIA PERSON (TV/RADIO) 6
OTHER, SPECIFY 7 O
DK8 X
RF 9 X
«GOINF_01»
«GOINF_02»
«GOINF_03»
«GOINF_04»
«GOINF_05»
«GOINF_06»
«GOINF_07»
«0_GOINF»
TIME PERIOD: 2
What type of information did you try to find? PROBE: Anything else?
INFORMATION ABOUT WHAT HAD HAPPENED 1
INSTRUCTIONS TO LEAVE 2
INSTRUCTIONS TO STAY 3
OTHER, SPECIFY 7 O
DK8 X
RF 9 X
«WHAI2_01»
«WHAI2_02»
«WHAI2_03»
«WHAI2_04»
«0_WHAI2»
TIME PERIOD: 2
How did you get this information? PROBE: Any other way?
FACE TO FACE 1
TELEPHONE 2
EMAIL/BLACKBERRY 3
PA ANNOUNCMENT 4
NIST NCSTAR 1-7, WTC Investigation 1 87
Appendix A
TV/RADIO 5
OTHER, SPECIFY 7 O
DK8 X
RF 9 X
«HOWG2_01»
«HOWG2_02»
«HOWG2_03»
«HOWG2_04»
«HOWG2_05»
«HOWG2_06»
«0_HOWG2»
TIME PERIOD: 2
And during the time between when you first became aware that something had happened at the World
Trade Center and when you first entered the stairwell or elevator to leave the tower, did you believe that
other people were in danger of being killed? AFTER BECOMING AWARE OF THE EVENT, BUT
BEFORE EVACUATION
=> YRIS2
ifORISK=l
YES 1
NO 2
DK 8
RF 9
«ORIS2»
TIME PERIOD: 2
During that time period, did you believe you were in danger of being killed?
=> PEODO
ifYRISK=l
YES 1
NO 2
DK 8
RF 9
«YRIS2»
TIME PERIOD: 2
1 88 NIST NCSTAR 1-7, WTC Investigation
Telephone Interview Script
During this time period, what were the people around you doing? PROBE: Were they doing anything
else? AFTER BECOMING AWARE OF THE EVENT, BUT BEFORE EVACUATION
$E0 10
NOONE AROUND/WAS ALONE 00 X
TALKING TO OTHERS 1
GATHERING PERSONAL/WORK ITEMS 02
SEARCHING FOR OTHERS 03
CALLING OTHERS 04
FIGHTING FIRE/SMOKE 05
LOCKING UP 06
WORKING 07
EVACUATING THE TOWER 08
CRYING, RUNNING AROUND, IN SHOCK 09
HELPING OTHERS 10
OTHER, SPECIFY 97 O
DK98 X
RF 99 X
«PEODO_01»
«PEODO_02»
«PEODO_03»
«PEODO_04»
«PEODO_05»
«PEODO_06»
«PEODO_07»
«PEODO_08»
«PEODO_09»
«PEODO_10»
«0_PEODO»
TIME PERIOD: 2
Did the people around you start evacuating before you did?
=> DOBEF
ifPEODO=08
YES 1
NO 2
NIST NCSTAR 1-7, WTC Investigation 1 89
Appendix A
DK 8
RF 9
«EVACB»
TIME PERIOD: 2
Did you do any of the following before starting your evacuation?
$E 1 9
Talk to another person face to face 1
Gather personal items 02
Telephone other people 03
Continue working 04
Save or transfer computer files 05
Search for others 06
Fight fire or smoke 07
Move to another floor 08
Help others 09
Logging off/shutting down computer 10
NONE OF THESE 11 X
«DOBEF_01»
«DOBEF_02»
«DOBEF_03»
«DOBEF_04»
«DOBEF_05»
«DOBEF_06»
«DOBEF_07»
«DOBEF_08»
TIME PERIOD: 2
Did you do anything else during this time?
ENTER RESPONSE 1 O
NO OTHER ACTIVITIES
DK8
RF 9
«OACTI»
«0 OACTI»
190
NIST NCSTAR 1-7, WTC Investigation
Telephone Interview Script
TIME PERIOD: 2
Before you began your evacuation, was there anything you wanted to do, but couldn't?
YES 1
NO 2 => SEE01
DK 8 => SEE01
RF 9 => SEE01
«WANTD»
TIME PERIOD: 2
What was that? PROBE: Anything else?
$E 1 7
GATHER WORK ITEMS 01
GATHER PERSONAL BELONGINGS 02
CALL FRIEND/FAMILY MEMBER 03
FIND FRIEND/COWORKER 04
HELP FRIEND/COWORKER 05
LOCKUP 06
EVACUATE IMMEDIATELY 07
OTHER, SPECIFY 97 O
DK98 X
RF 99 X
«WANAC_01»
«WANAC_02»
«WANAC_03»
«WANAC_04»
«WANAC_05»
«WANAC_06»
«WANAC_07»
«WANAC_08»
«0_WANAC»
TIME PERIOD: 2
Why couldn't you do that/those things?
$E 1 9
NIST NCSTAR 1-7, WTC Investigation 191
Appendix A
AFRAID 01
LOCKED DOORS 02
PHONE LINES DEAD 03
INJURED 04
EXIT BLOCKED 05
TOO CROWDED 06
TOLD TO STAY IN BUILDING
TOLD TO LEAVE 08
FATIGUE 09
DISABLED 10
SMOKE 1 1
DAMAGE TO FLOOR 12
WAS HELPING OTHERS 13
OTHER, SPECIFY 97 O
DK98
RF 99
«WHYNO_01»
«WHYNO_02>:
«WHYNO_03>:
«WHYNO_04>:
«WHYNO_05>:
«WHYNO_06>:
«WHYNO_07>:
«WHYNO_08>:
«WHYNO_09>:
«WHYNO_10>:
«WHYNO_ll>:
«WHYNO_12>:
«WHYNO_13>:
«WHYNO_14>:
«0 WHYNO»
07
192
NIST NCSTAR 1-7, WTC Investigation
Telephone Interview Script
Still thinking about the time between when you first became aware that something had happened at the
World Trade Center and when you entered the stairwell or elevator to leave the tower, did you notice any
of the following? FOLLOW UP: Was that in your immediate area or outside the Tower?
Did Not Notice
Noticed in Immediate Area
Noticed Outside the
Tower
Smoke
Fire or Flames
Fireballs
Collapsed walls
Jet Fuel
Severely or fatally injured
people
Sprinklers going on
A fire alarm sounding
Power outage or flickering lights
Fallen ceiling tiles
Extreme heat
«SEE01_01»
«SEE01_02»
TIME PERIOD: 2
Were there any disaster related events going on around you at this time?
=> EVACF
ifOR[SEE01-SEEll]=2-3
YES 1
NO 2 => HELPY
DK 8 => HELPY
RF 9 => HELPY
«ODISE»
TIME PERIOD: 2
What was going on?
ENTER RESPONSE 1 O
DK8
RF 9
«GOIN2»
«0 GOIN2»
NISTNCSTAR 1-7, WTC Investigation
193
Appendix A
TIME PERIOD: 2
Were you still on the<WHFL2>floor at this time? RANGE: 1st - 1 10th FLOOR IF YES,
SELECT/ENTER FLOOR IF NO, ASK WHICH FLOOR THEY WERE ON AND SELECT/ENTER IT
$E 1 110
=>+l
if (AND[SEE01-SEE11]=1) AND PEODO>0 AND PEODO<98
BASEMENT 990
CONCOURSE/LOBBY 991
PLAZA 992
ELEVATOR 993
OTHER, SPECIFY 997 O
DK998
RF 999
«EVACF»
«0_EVACF»
TIME PERIOD: 2
Did anyone help you in any way before you started your evacuation?
YES 1
NO 2 => DECID
DK 8 => DECID
RF 9 => DECID
«HELPY»
TIME PERIOD: 2
Who helped you? PROBE: Anyone else? WE WANT THEIR ROLE NOT THE NAME OF THE
PERSON
POLICE OFFICER/FIREFIGHTER 1
COWORKER 2
STRANGER 3
FLOOR WARDEN 4
MANAGER/SUPERVISOR 5
OTHER, SPECIFY 7 O
DK8 X
1 94 NIST NCSTAR 1-7, WTC Investigation
Telephone Interview Script
RF 9 X
«WHOHE_01»
«WHOHE_02»
«WHOHE_03»
«WHOHE_04»
«WHOHE_05»
«WHOHE_06»
«0_WHOHE»
TIME PERIOD: 2
What did they help you with? PROBE: Anything else?
$E 1 7
LOCATING OTHERS 01
HELPING OTHERS 02
FINDING EXITS 03
TREATING YOUR INJURIES 04
PROVIDED INFORMATION/INSTRUCTIONS 05
GATHER BELONGINGS 06
CALM DOWN/EMOTIONAL ASSISTANCE 07
OTHER, SPECIFY 97 O
DK98 X
RF 99 X
«WHATD_01»
«WHATD_02»
«WHATD_03»
«WHATD_04»
«WHATD_05»
«WHATD_06»
«WHATD_07»
«WHATD_08»
«0_WHATD»
TIME PERIOD: 2
What was the one thing that made you decide to evacuate?
WAS TOLD TO EVACUATE 1
NIST NCSTAR 1-7, WTC Investigation 1 95
Appendix A
FRIENDS CO-WORKERS EVACUATED
AFRAID/FELT IN DANGER 3
FIRE ALARM WAS GOING OFF 4
SAW SMOKE 5
SAW FIRE 6
OTHER, SPECIFY 7 O
DK8
RF 9
«DECID»
«0 DECID»
How many minutes had passed before you started to evacuate? IF NEEDED: How much time passed
between when you first became aware that something had happened to the World Trade Center and when
you entered the stairwell or elevator to leave the tower. THIS IS NOT TIME TO EVACUATE.
PLEASE CLARIFY WITH RESPONDENT IF TIME APPEARS TOO LONG. RESPONDENT WAS
IN<WHTW2> RANGE FOR TOWER 1: 1 - 103 MINUTES RANGE FOR TOWER 2: 1 - 75
MINUTES
$E 1 103
DK 998
RF 999
«TIMEP»
SKIP FOR TOWERS
=> EVAC2
Else=>+1
ifWHTW2=2
«SKIP1»
Did you begin your evacuation. . . WE ARE INTERESTED IN WHAT THEY KNOW NOW. THEY
MAY NOT HAVE KNOWN WHEN THEY WERE EVACUATING, BUT NOW THEY CAN TELL US
WHEN IT WAS.
Before the plane hit Tower 2 1
After the plane hit Tower 2, but before the Tower 2 collaspe 2
After the Tower 2 collaspe 3
DK 8
RF 9
«EVAC1»
1 96 NIST NCSTAR 1 - 7, WTC Investigation
Telephone Interview Script
SELECT1
$S NS=2 C0=1 IN=EVAC1<=1 ;C0=2 IN=EVAC1<=2 ;
Before the plane hit Tower 2 1
After the plane hit Tower 2, but before the Tower 2 collaspe 2
After the Tower 2 collaspe 3
DK8
RF 9
«SEL1»
SELECT2
Before the plane hit Tower 2 1
After the plane hit Tower 2, but before the Tower 2 collaspe 2
After the Tower 2 collaspe 3
DK8
RF 9
«SEL2»
Did you begin your evacuation. . .
=> EVCSO
ifEVACl>0
Before the plane hit Tower 2 1
After the plane hit Tower 2 2
DK8
RF 9
«EVAC2»
SELECT4
$S CO=l IN=EVAC2<=1 ;
Before the plane hit Tower 2 1
After the plane hit Tower 2 2
DK8
RF 9
«SEL3»
NIST NCSTAR 1-7, WTC Investigation 1 97
Appendix A
Was there anything that kept you from evacuating sooner?
YES, RECORD RESPONSE 1 O
NO 2
DK8
RF 9
«EVCSO»
«0_EVCSO»
TIME PERIOD: 3
When you began your evacuation, were you alone or with other people? PEOPLE THAT THEY KNOW,
PEOPLE THAT THEY WERE TALKING WITH
ALONE 1
WITH OTHER PEOPLE 2
DK8
RF 9
«ALONE»
TIME PERIOD: 3
Which stairwell did you use for your evacuation?
STAIRWELL A
1
STAIRWELL B
2
STAIRWELL C
3
USED ELEVATOR
4
=> FOLA1
OTHER, SPECIFY
7
DK8 X
RF 9 X
«STAIR_01»
«STAIR_02»
«STAIR_03»
«STAIR_04»
«STAIR_05»
«0 STAIR»
TIME PERIOD: 3
Which side of the building was the stairwell located on?
198 NIST NCSTAR 1 - 7, WTC Investigation
Telephone Interview Script
=> /WHYST
if NOT STAIR=8,7
NORTH 1
SOUTH 2
EAST 3
WEST 4
OTHER, SPECIFY 7 O
DK8
RF 9
«WHISI»
«0_WHISI»
TIME PERIOD: 3
Why did you choose that/those stairwell(s) for your evacuation? PROBE: Any other reason?
CLOSEST ONE 1
FOLLOWED OTHER PEOPLE TO IT 2
OTHER EXITS WERE BLOCKED 3
SAME AS I USED IN PREVIOUS EMERGENCY 4
I WAS TOLD TO USE THIS STAIRWELL 5
OTHER, SPECIFY 7 O
DK8 X
RF 9 X
«WHYST_01»
«WHYST_02»
«WHYST_03»
«WHYST_04»
«WHYST_05»
«WHYST_06»
«0_WHYST»
TIME PERIOD: 3
At any time during your evacuation, did you leave that/those stairwell(s)? DO NOT INCLUDE PEOPLE
WHO FOLLOWED THE PASSAGE WHERE THE STAIRWELLS START AND END.
YES 1
NO 2 => FOLA1
NIST NCSTAR 1-7, WTC Investigation 1 99
Appendix A
DK8 =>F0LA1
RF 9 => F0LA1
«LEVST»
TIME PERIOD: 3
Which floor were you on when you left the stairwell? IF RESPONDENT UNSURE, SELECT 997 AND
RECORD RANGE OF FLOORS EXAMPLE: 34-40
$R1 110
UNSURE, RECORD RESPONSE 997 O
«FLLST»
«0_FLLST»
TIME PERIOD: 3
Why did you leave the stairwell? PROBE: Any other reason?
$E 1 9
I GOT LOST 01
WAS TOLD TO LEAVE STAIRWELL 02
TO HELP SOMEONE 03
TO GO BACK AND GET SOMETHING 04
TOO CROWDED 05
SMOKE IN STAIRWELL 06
PATH OBSTRUCTED 07
A LOCKED DOOR 08
STAIRWELL LED TO A FLOOR 09
OTHER, SPECIFY 97 O
DK98
RF 99
«WHYLS_01»
«WHYLS_02»
«WHYLS_03»
«WHYLS_04»
«WHYLS_05»
«WHYLS_06»
«WHYLS_07»
«WHYLS_08»
200 NIST NCSTAR 1 - 7, WTC Investigation
Telephone Interview Script
«WHYLS_09»
«WHYLS_10»
«0_WHYLS»
Screen [Template 3] -> FL0A5
=>+l
ifFLWAR>l
Did any of the following help you evacuate while you were in the building?
Yes No DK RF
Instructions or assistance from your floor warden
Instructions or assistance from Police or Firefighters
Support and encouragement from others
Exit signs
Photo luminescent paint in stairwells
«FOLAl»
Screen [Template 3] -> EVCM7
=>+l
if NOT STAIR<4
Did any of the following make your evacuation more difficult while you were in the building?
Yes
No
DK
RF
Crowded stairwells
Firefighters or Police moving up stairwell
Disabled or injured people being taken down stairwell
Locked doors
Poor lighting
Confusing or missing signs
Lack of clear instructions
«EVCM1»
Screen [Template 3] -> EXP1 1
NISTNCSTAR 1-7, WTC Investigation
201
Appendix A
Please tell me if you noticed any of the following at any time during your evacuation. FOLLOW UP:
Was that in your immediate area or outside the Tower?
Did Not Notice
Noticed in Immediate Area
Noticed Outside the Tower
Smoke
Fire or Flames
Fireballs
Collapsed walls
Jet Fuel
Severely or fatally injured people
Sprinklers going on
A fire alarm sounding
Power outage or flickering lights
Fallen ceiling tiles
Extreme heat
«EXP01_01»
«EXP01_02»
TIME PERIOD: 3
During your evacuation, did you turn back at any time? "TURN BACK" MEANS "GO BACK UP".
YES 1
NO 2 => EXITS
DK 8 => EXITS
RF 9 => EXITS
«TURNB»
TIME PERIOD: 3
Why did you turn back? PROBE: Any other reason?
$E 1 7
I GOT LOST 01
I WAS TOLD TO TURN BACK 02
TO HELP SOMEONE 03
TO GET SOMETHING 04
IT WAS TOO CROWDED 05
SMOKE IN THE STAIRWELL 06
202
NIST NCSTAR 1-7, WTC Investigation
Telephone Interview Script
MY PATH WAS OBSTRUCTED 07
OTHER, SPECIFY 97 O
DK98 X
RF 99 X
«WHYTB_01»
«WHYTB_02»
«WHYTB_03»
«WHYTB_04»
«WHYTB_05»
«WHYTB_06»
«WHYTB_07»
«WHYTB_08»
«0_WHYTB»
TIME PERIOD: 3
Did you exit the stairwell or elevator to the mezzanine or to the concourse?
MEZZANINE 1
CONCOURSE 2
OTHER, SPECIFY 7 O
DK8
RF 9
«EXITS»
«0_EXITS»
TIME PERIOD: 3
How much time passed between the moment you first began your evacuation to when you exited the
Tower? PLEASE CLARIFY WITH RESPONDENT IF TIME APPEARS TOO LONG. RESPONDENT
WAS IN<WHTW2> RANGE FOR TOWER 1: 1 - 103 MINUTES RANGE FOR TOWER 2: 1 - 75
MINUTES
$E 1 103
DK 998
RF 999
«TIMP2»
SKIP FOR TOWERS
=>+2
NIST NCSTAR 1-7, WTC Investigation 203
Appendix A
Else=>+1
ifWHTW2=2
«SKIP2»
TIME PERIOD: 3
Did you exit the tower. . .
Eliminate -> 2
According to NOT SEL1-SEL2
Before the plane hit Tower 2 1 => GETOU
After the plane hit Tower 2 but before the Tower 2 collapse, or 2 => GETOU
After the Tower 2 collapse 3 => GETOU
DK 8 => GETOU
RF 9 => GETOU
«EXIT1»
TIME PERIOD: 3
Did you exit the tower. . .
Eliminate -> 1
According to NOT SEL3
Before the plane hit Tower 2, or 1
After the plane hit Tower 2 2
DK8
RF 9
«EXIT2»
Please remember that this study is intended as a fact finding mission and not a fault finding mission. It is
crucial that we determine why some people were successful in their evacuation while others were not.
Was there anyone on your floor that was not successful in their evacuation?
YES 1
NO 2
=> PHYSI
DK 8
=> PHYSI
RF 9
=> PHYSI
«GETOU»
Why didn't they make it out? PROBE: Any other reason?
204 NIST NCSTAR 1-7, WTC Investigation
Telephone Interview Script
$E 1 8
WAS INJURED 01
WAS DISABLED 02
REFUSED TO LEAVE 03
DID NOT THINK IT WAS SERIOUS 04
STAYED BACK TO HELP SOMEONE 05
WAS TOLD TO STAY 06
STRUCTURAL DAMAGE 07
SMOKE OR FIRE 08
OTHER, SPECIFY 97 O
DK98 X
RF 99 X
«WHYNG_01»
«WHYNG_02»
«WHYNG_03»
«WHYNG_04»
«WHYNG_05»
«WHYNG_06»
«WHYNG_07»
«WHYNG_08»
«WHYNG_09»
«0 WHYNG»
On September 1 1th, 2001, did you have any physical problems that made it more difficult for you to leave
the tower? Please do not include injuries caused by the incident or evacuation.
YES 1
NO 2
=>AGE
DK 8
=>AGE
RF 9
=>AGE
«PHYSI»
What type of physical problem? PROBE: Anything else?
$E 1 9
BLIND/PARTIALLY BLIND 01
DEAF 02
NIST NCSTAR 1-7, WTC Investigation 205
Appendix A
IN WHEELCHAIR 03
NEED WALKING ASSISTANCE
OBESITY 05
HEART CONDITION 06
PREGNANT 07
ASTHMA 08
ELDERLY 09
OTHER, SPECIFY 97 O
DK98 X
RF 99 X
«LIMIT_01»
«LIMIT_02»
«LIMIT_03»
«LIMIT_04»
«LIMIT_05»
«LIMIT_06»
«LIMIT_07»
«LIMIT_08»
«LIMIT_09»
«LIMIT_10»
«0 LIMIT»
04
What is your age? RANGE: 1 - 98 YEARS
$E 1 99
RF 99
«AGE»
READ ONLY IF YOU CAN'T TELL. What is your gender?
MALE 1
FEMALE 2
RF 9
«GEND»
What language do you speak best?
ENGLISH 1
206
NIST NCSTAR 1-7, WTC Investigation
Telephone Interview Script
SPANISH 2
OTHER, SPECIFY 7 O
DK8
RF 9
«PLANG»
«0_PLANG»
Were you working in Tower 1 or Tower 2 during the 1993 bombing?
=>SAY11
ifYRWRK>1993
YES 1
NO 2 => CONCR
DK 8 => CONCR
RF 9 => CONCR
«WBOMB»
During the 1 993 bombing, did you evacuate immediately or wait to evacuate?
EVACUATE IMMEDIATELY 1
WAIT TO EVACUATE 2
DK 8 => +2
RF 9 => +2
«EVBOM»
At the time of the 1993 bombing, did you feel you that your decision to<EVBOM>was the right decision?
YES 1
NO 2
DK 8
RF 9
«DEC93»
After the 1993 bombing how concerned were you that terrorists would attack the World Trade Center?
Were you...
Extremely Concerned 1
Very Concerned 2
Moderately Concerned 3
NIST NCSTAR 1-7, WTC Investigation 207
Appendix A
Slightly Concerned 4
Not at all Concerned 5
DK8
RF 9
«CONCR»
Is there anything else you would like to say about your experience on September 1 1th?
YES, RECORD RESPONSE 1 O
NO 2
DK 8
RF 9
«SAY11»
«0_SAY11»
IMPACT FLOOR FLAG
=> *
if IF(((WHTW2=1 AND WHFL2>91 AND WHFL2<99) OR (WHTW2=2 AND WHFL2>77 AND
WHFL2<111)),1,0)
IMPACT FLOOR FLAG 1
«FFLAG»
163: LFLAG
Single
min = 1 max =1 1=1
2003/09/18 15:21
LOCATION FLAG
=> *
if IF((WHFL2>990 AND WHFLK994), 1 ,0)
LOCATION FLAG 1
«LFLAG»
EVENT FLAG
=> *
if IF(((AND[NOT02-NOT06]=2-3) OR (AND[SEE02-SEE06]=2-3) OR (AND[EXP02-EXP06]=2-
3)), 1,0)
208 NIST NCSTAR 1 - 7, WTC Investigation
Telephone Interview Script
EVENT FLAG 1
«EFLAG»
DISABILITY FLAG
=> *
ifIF((PHYSI=l),l,0)
DISABILITY FLAG 1
«DFLAG»
ROLE FLAG
=> *
ifIF((ROLES=l-4),l,0)
ROLE FLAG 1
«RFLAG»
We may be interested in learning more about your experience on September 1 1th. Would it be okay if we
follow up with you sometime in the future to get more detailed information on your evacuation
experience?
=>+l
ifFFLAG+LFLAG+EFLAG+DFLAG+RFLAG==0
YES 1
NO 2
«FOLUP»
PRESS ENTER TO CONTINUE
Those are all the questions we have. The valuable information you provided will help designers and
engineers improve building safety, and help emergency planners improve building evacuation procedures.
Thank you so much for taking the time to talk with me, and have a good day/evening. Good-bye.
END OF SURVEY 1 D => /INT99
«THANK»
NISTNCSTAR 1-7, WTC Investigation 209
Appendix A
This page intentionally left blank.
210 NIST NCSTAR 1-7, WTC Investigation
Appendix B
Qualitative Analysis Coding
The qualitative data from the face-to-face interviews was used to enrich the statistically-based telephone
interview data by providing detailed descriptions of the experiences of occupants of the two towers on
September 1 1 , 200 1 . These data identified unknown information, evaluated technical hypotheses, and
explored conscious and subconscious motivations for occupant behaviors, while allowing for
comparisons to the telephone interview data. This appendix provides an annotated listing of the coding
used by National Institute Standards and Technology (NIST) to analyze the face-to-face interview data. It
was intended to serve the dual purposes of the project by identifying 1) building damage and
environmental conditions before or during evacuation and 2) patterns of behavior before or during
evacuation that may have helped or hindered occupants' evacuation. Some of the codes would apply to
an entire interview (for example, the tower 1 , tower 2, or Building 7 code simply indicates the building
where the respondent was located at the beginning of the event), some part of their evacuation experience
(such as the counterflow, crowded, or elevator codes), or an observation of their surroundings (such as the
floor damage, building damage, saw fire, or saw smoke codes). The codes are simply presented in
alphabetical order.
Alarm: Heard fire or other alarm (not public address system)
Alternative Activity: Decision to suspend evacuation in favor of another activity (call, get
drink/food/possession ...)
Antisocial Behavior: Any directly observed behavior which presents obstacle to people getting out
Assist: Assisted or motivated an injured / handicapped / reluctant person
Bldg Employee: Building service employee (janitor, elevator operator, contract service employee — not
building security) as an information source
Boss or higher: Boss, manager, or higher as an information source
Building 7: Occupant was in Building 7 when Tower 1 was struck
Building Damage: Directly observed damage to building structure after beginning evacuation (see floor
damage for damage on floor where they began their evacuation)
Building Security: Uniformed building security personnel (not Port Authority, NYPD, or FDNY)
Cellular Phone: Use of a cellular phone as an information source
Concourse: Occupant mentions being in the Concourse during their evacuation. Autocoded with phrases
such as Concourse Borders shopping mall,...
Confusion: Unaware of where to go because of a lack of visual certainty
Counterflow: Movement of people or responders against the flow of egress delayed evacuation
Coworker: Coworker as an information source
Crowded: Egress slowed because of density, but no specific identifiable cause
Elevator - Considered: Occupant considered using elevators but did not make use of elevators for
evacuation. Code should also be used when occupant states they knew not to use elevators.
NIST NCSTAR 1-7, WTC Investigation 211
Appendix B
Elevator - Used: Used an elevator during evacuation
Email: Used email or PDA device as an activity
Email/PDA: Use of email or PDA as an information source
Evacuation Decision: Marks when respondent decided they needed to leave. Note that his may occur
more than once if the occupant decides to interrupt their evacuation on it has begun.
Experience Aid: Prior experience aided evacuation
Experience Obstacle: Experience caused them to delay evacuation
FBI: Occupant specifically mentioned seeing someone identified with the Federal Bureau of Investigation
FDNY: Fire department personnel as an information source
Fire Affecting Egress: Observed fire in egress path
Firefighting by Occupants: Occupant engaged in firefighting activities prior to or during evacuation
First Awareness: First indication that respondent became aware that something was wrong on the
morning of September 1 1
Floor Damage: Directly observed damage to building structure on floor where they began evacuation
prior and to beginning evacuation (see building damage for damage during evacuation and not on
floor where they began their evacuation)
Floor Start: Starting floor for evacuation that is different from the employer assigned floor
Floors 1 to 9: Occupant began their evacuation in the stated range of floors
Floors 10 to 19: Occupant began their evacuation in the stated range of floors
Floors 100 to 110: Occupant began their evacuation in the stated range of floors
Floors 20 to 29: Occupant began their evacuation in the stated range of floors
Floors 30 to 39: Occupant began their evacuation in the stated range of floors
Floors 40 to 49: Occupant began their evacuation in the stated range of floors
Floors 50 to 59: Occupant began their evacuation in the stated range of floors
Floors 60 to 69: Occupant began their evacuation in the stated range of floors
Floors 70 to79: Occupant began their evacuation in the stated range of floors
Floors 80 to 89: Occupant began their evacuation in the stated range of floors
Floors 90 to 99: Occupant began their evacuation in the stated range of floors
Floors below ground: Occupant began their evacuation in the stated range of floors
Followed Crowd: Respondent reports following a crowd, or else avoiding a crowd, eg a stair looks too
crowded.
Fruin: An indication of Fruin level of service for crowding in the stairwells
Get Information: Looking for more information (not milling)
Get Out Instruction: Identifying the instruction to evacuate the building - it is of interest as to who this
instruction came from (secretary or supervisor, for instance).
Health: Occupant reports being tired, sick, in pain ... effecting egress
212 NIST NCSTAR 1-7, WTC Investigation
Qualitative Analysis Coding
Helping: Stopped or slowed by helping other who needed assistance
High: Floors 77 to 1 10 in the Towers and 26 to 47 in Building 7
Immediate Evacuation: Where occupant made an immediate decision to evacuate
Instruction: Instruction to take action counter to continuing or commencing evacuation (any source).
Instruction Aid: Received an instruction (any source) that aided evacuation
Interesting behavior: This code is a chance to identify any interesting behavior to highlight for the
report.
Internal Aids: Any thoughts or techniques which calmed occupant during evacuation (counting,
humming, thinking of family ...)
Know 1993: While not there, occupant was influenced by second-hand knowledge of the 1993
WTC bombing
Left Stairwell: Occupant left stairwell during evacuation
Location: Indicates a specific location within a building before or during evacuation
Location - Concourse: Indicates the occupant was on the Concourse Level
Locked/ Jammed Door: First hand observation of a locked or jammed door effecting egress
Low: Floors 1 to 40 in the Towers and 1 to 25 in Building 7
Media: Use of the media (TV, radio ...) as an information source
Medical Disability: Asthma, illness, pregnancy
Medium: Floors 43 to 74 in the Towers
Mill: Talked with other people inside the building as an activity
Milling Aid: Milling as an aid to evacuation
No alarm: Occupant specifically reports not hearing an alarm at any time during their evacuation
Other Activity: Activities of interest to the NIST investigation not otherwise categorized
Other Egress Aid: Aid to egress of interest to the NIST investigation not otherwise categorized
Other Info: Other information source of interest to the NIST investigation not otherwise categorized
Other Mobility Challenge: Mobility-impaired information of interest to the NIST investigation not
otherwise categorized
Other Obstacle: Obstacle effecting egress not otherwise categorized
Overweight: Overweight or out of shape
Pager: Occupant reported using a pager at some time before or during their evacuation
Phone: Used a phone to receive information (not cellular phone)
Phone Call: Respondent made use of a telephone as an activity (not cell phone). Typically to inform
someone of status ~ family, 911, etc. For phone as an information source, use cellular phone or
phone
Photoluminscent Paint: Photoluminescent paint mentioned as an aid to evacuation
Physical Assistance: Received physical assistance which aided evacuation
NIST NCSTAR 1-7, WTC Investigation 213
Appendix B
Physical Disability: Wheelchair, crutches, cane, bad knees, ...
Police: NYPD or other official gun-carrying officers as an information source
Port Authority: PAPD or person with building responsibility as an information source
Practice Drill: Participated in practice drills at WTC
Present 1993: Occupant was present at the 1993 WTC bombing
Prior Evac Experience: Participated in previous building evacuation (WTC or not), except for the 1993
bombing.
Prosocial Behavior: Letting people in line, crowd calming conversations, singing, counting out loud,
emergent leadership or other behaviors which aid evacuation
Public Address Aid: Information from public address system assisted evacuation
Public Address System: Public address system as an information source
Quote: A short quotation that may warrant inclusion in the final report because it creates, supports, or
refutes a hypothesis or is particularly compelling account of a particular event
Rest on stairs: This code identifies when the respondent actually rests in the stairs and/or observes this
behavior.
Retrieve: Retrieved a personal or business belonging
Risk - High Level: High perception of danger faced by the occupant
Risk - Low Level: Low perception of the danger faced by respondent
Saw Airplane: Directly observed airplane hitting occupant's building (saw other impact for observation
of plane hitting other building)
Saw FDNY: Respondent mentions the presence of firefighters during their egress whether or not they
helped or hindered egress or gave instructions
Saw Fire: Directly observed flames within the building before or during evacuation (does not include
observation of fire on exterior of another building)
Saw Injured: Directly observed an injured or dead person, not including themselves
Saw Other: Saw something relevant to NIST investigation not otherwise categorized
Saw Other Impact: Directly observed impact of plane into a building other that the one the respondent
was in (for example someone in T2 seeing plane hit Tl)
Saw People: Observed someone doing something noteworthy, but it did not impact egress
Saw Smoke: Directly observed smoke within the building before or during evacuation (does not include
observation of smoke on exterior of another building)
Saw Water: Directly observed water within the building before or during evacuation (not drinking water)
Saw/Felt/Heard Impact: Indication of the impact of a plane hitting the building. (If you saw the impact
of the plane hitting the other building and felt it in your building, it would get coded twice; once
with this code and once with Saw other impact)
Search: Searched for a person or people
Shoes: Poor footwear choice or other's footwear cluttering pathway effects egress
Signage: Signs posted in evacuation route aided evacuation
214 NIST NCSTAR 1-7, WTC Investigation
Qualitative Analysis Coding
Slow Occupant: A slow occupant ahead of respondent in line impeded egress
Smell/Saw Fuel: Direct observation or smell of jet fuel in the building
Smoke Affecting Egress: Observed smoke in egress path
Stair Unknown: Direct mention that respondent did not know which stairwell was used for evacuation
Stairway from basement: Occupant used stairs from below-ground levels during evacuation
Stairwell A: Occupant was in Stairwell A
Stairwell A or C : Occupant was in Stairwell A or C (narrow stairwell)
Stairwell B: Occupant was in Stairwell B
Stairwell C: Occupant was in Stairwell C
Strange: For something that just doesn't make sense and is worth further investigation
Superflow: Prosocial behavior of letting higher priority evacuees (i.e., injured) pass faster delayed
evacuation
Superflow Aid: By helping someone injured or handicapped, was able to egress faster than they
otherwise would have
T2 Collapse: Note by respondent of collapse of Tower 2
T2 hit: Note by respondent of Tower 2 being hit
Time: Indicates a relative or absolute time
Tower 1: Occupant was in Tower 1 when Tower 1 was struck
Tower 2: Occupant was in Tower 2 when Tower 1 struck
Training Aid: Prior training aided evacuation
Transfer Floor: Occupant notes use of a transfer floor during egress by stairwell
Trapped: Occupant was trapped prior to or during evacuation
Victim: Mention of people who died in the event
Visibility: low light condition or dust (but not smoke) effecting egress
Visitor: Respondent specifically mentions being a visitor or being with a visitor at some point during
their evacuation
Walkie-Talkie: Respondent mentions the use of a walkie-talkie either by the respondent or by someone
else
Warden: Activities associated with responsibilities of the floor wardens
Warden Info: Floor warden as a source of information
Water: Water within the building before or during evacuation (not drinking water) was an obstacle to
evacuation
NIST NCSTAR 1-7, WTC Investigation 215
Appendix B
This page intentionally left blank.
216 NIST NCSTAR 1-7, WTC Investigation
Appendix C
Causal Modeling
The telephone interview results permit rigorous statistical analysis of individual components (questions).
In order to understand what factors impacted the overall evacuation time of the average occupant in
World Trade Center (WTC) 1 or WTC 2, two primary dependent variables were predicted: how long an
individual delayed initiating their evacuation, where initiation is defined as entering a stairwell or elevator
with the intention of exiting the building; and how long an individual spent traversing the stairwells. The
sum of these two times was the total evacuation time. Multivariate regression modeling was utilized to
rigorously establish factors which contributed to increasing the overall evacuation time.
C.l PREDICTING EVACUATION DELAY IN THE WORLD TRADE CENTER
This project analyzed the factors (variables) and social processes (the major paths of causal links between
variables) that influenced people delaying the initiation of their evacuation out of WTC 1 and WTC 2 on
September 11, 2001. Evacuation delay was defined as the number of minutes that passed from when a
person first became aware that something was wrong until they began their evacuation.
Method
The purpose of the Telephone Interviews was to collect 800 computer assisted telephone interviews
(CATI) of persons occupying either of the two WTC towers (WTC 1 and WTC 2) at the time of the
terrorist attacks on September 11, 2001 in order to explore occupants' evacuation experiences. A total
sample size of 800, with an allocation of 400 to each tower, was determined to simultaneously maximize
the statistical precision within each tower. Estimates of percentages from tower specific survey data at
400 exhibit sampling errors not greater than 2.5 percent, and 98 percent confidence intervals of
percentages are no greater than ± 5 percent. This level of precision is more than adequate for examining
characteristics of occupants and egress attributes. Additionally, mutlivariate modeling requires the use of
F tests to determine significance of the regression models, i.e., testing the null hypothesis that the ratio of
explained variance to error/residual variance is equal to or less than zero. The sample size of 400 per
tower is more than adequate for this analysis since, for example, in a model featuring 20 independent
variables, a sample size of 400 and a .05 level of significance (Type I error), the power of the F test to
detect an R-squared of .06 is just over 81 percent.
Attempts were made to equally divide the respondents among WTC 1 and WTC 2 occupants (i.e., n=400
occupant telephone interviews from each tower). Within each of the WTC buildings, independent
proportionate stratified samples of survivors were drawn. In other words, each occupant of a particular
tower had an equal probability of being selected. The sample was collected and weighted in accordance
with the analysis presented in Section 3.2.2 of this report. The questions posed to the respondents are
included in earlier in this report.
NIST NCSTAR 1-7, WTC Investigation 217
Appendix C
Pre Modeling Analysis
Two bodies of research and theory were reviewed to inform our consideration of what explanatory
factors/variables should be used to predict occupant evacuation from the towers.
First, we compiled a comprehensive list of the 300 publications in the research record on human
response/evacuation to risk events/information for natural and technological hazards and disasters, for
example, research on floods, hurricanes, earthquake predictions, tornadoes, nuclear power plant accidents,
hazardous chemical spills and many others. These studies included not only examinations of evacuations,
but also events in which other protective actions were warranted. This literature was reviewed,
summarized, and synthesized for salient theoretical constructs and the relationships between them. This
bibliography is available in both standard and annotated formats from the Natural Hazards Research and
Applications Information Center at the University of Colorado at Boulder.
Second, we accessed the record of research on human evacuation during high rises fires. Experts in both
areas of research developed a list of the most salient factors to include in this segment of our research.
A range of key possible explanatory variables on which to gather data to use to predict evacuation delay
in towers 1 and 2 were selected. We collected telephone interview data on these factors, and then
statistically examined their import in predicting evacuation delay time. The list of factors follows.
Location in Building
Social Context
Demographic Characteristics
Pre-event Experience
Roles of Responsibility
Preparedness and Training
Environmental Cues
Social Cues
Receiving Information about the Event
Perceived Risk
Seeking Additional Information
Injuries
Obtaining Help from Others
Pre-evacuation Activities
218 NIST NCSTAR 1 - 7, WTC Investigation
Causal Modeling
Preliminary statistical analyses were performed to determine which of these had and did not have salience
in predicting evacuation delay out of WTC 1 and WTC 2. Those that had no significant impact on
evacuation delay were excluded from further consideration. The factors that had salient impacts on
evacuation delay were carried forward into the analysis to model tower evacuation.
The Model That Was Estimated
The model used to predict evacuation delay in WTC 1 and WTC 2 of the World Trade Center on
September 11, 2001 used factors that preliminary analyses suggested as salient, closely followed general
evacuation theory from the social and fire sciences, and is diagrammatically illustrated in Figure C-l.
Note that e 4 - e 7 are the unexplained variance for variables 4-7, respectively. The model can be
described as follows: (1) delay in evacuation initiation is a direct consequence of environmental cues,
floor, obtaining information without seeking it, perceived risk, seeking additional information, and taking
pre-evacuation actions; (2) taking pre-evacuation actions is a direct consequence of environmental cues,
floor, obtaining information without seeking it, perceived risk, and seeking additional information; (3)
seeking additional information is a direct consequence of environmental cues, floor, obtaining
information that was not sought, and perceived risk; and, finally, (4) perceived risk is a direct
consequence of environmental cues, floor, and obtaining information without seeking it. This model is
parsimonious, consistent with the evacuation theory that stems from research on existing evacuation and
risk communication research, and it well-represented the positive findings of our preliminary analyses of
the many variables that could have impacted evacuation delay.
Measurement
Environmental cues (Xi) was measured by asking respondents about the number and type of
environmental cues (severe signs of danger, and non-severe signs of danger) that they saw prior to
initiating there evacuation. Answers to these questions were coded as dummy variables (0 or 1) and then
added to form an environmental cues scale that could vary between and 2.
Floor (X 2 ) was measured by asking respondents what floor they were on when the event started.
Responses were coded as -7 (basement floors) to 105. Negative floor values were transposed into positive
ones since this measure sought to determine how many floors people were from their building's floor of
exit. Missing data was coded to the mean.
Obtaining information without seeking it (X 3 ) was measured by asking respondents: "Now please think
about the time period between when you first became aware that something had happened and when you
first entered a stairwell or elevator to leave the tower. During this entire time period were you given any
additional information about what was going on?" Answers were coded as a dummy variable where
1 = yes and = no or missing data.
Perceived risk (X 4 ) was measured by asking whether the occupant whether they believed that other people
were in danger of being killed during the time when they first became aware that something had happened
and when they first entered the stairwell or elevator to leave. Once again, answers were coded as a
dummy variable, where 1 = yes and = no or missing data. This measure of perceived risk was used
instead of danger to self of being killed because the latter contained insufficient variance to include in the
analysis.
NIST NCSTAR 1-7, WTC Investigation 219
Appendix C
Perceived
Risk
s
/
Pre-Evacution
Actions
(x.)
Environmental
Cues
(x,)
/
Obtained
Information
(x,)
Delay
Initiating
Evacuation
(x 7 )
Sought
Information
w
\
Figure C-l. Model of evacuation delay.
Seeking additional information (X 5 ) was measured by asking the respondent whether they tried to get
additional information about what was going on. Answers were coded a 1 = yes (for both successful and
unsuccessful attempts to get additional information), and = no.
Taking pre-evacuation actions (X 6 ) was measured by asking respondents whether or not they performed
any of the following tasks: talk to another person face-to-face, gather personal items, telephone other
people, continue working, save or transfer computer files, search for others, fight fire or smoke, move to
another floor, or help others. These nine items were added to create a scale of taking actions (post- first
became aware that something was wrong but pre-evacuation initiation) that varied between and 9.
Finally, respondents were asked about evacuation delay (X 7 ) that is how many minutes passed from the
time when they first became aware that something was wrong until they actually began their evacuation.
Their responses were coded as an interval scale of minutes that varied between 1 to 80 minutes for
WTC 1, and 1 to 75 minutes for WTC 2. The means, medians, and modes for WTC 1 were 5.61 minutes,
220
NIST NCSTAR 1-7, WTC Investigation
Causal Modeling
3.00 minutes, and 1 minute. These same values for WTC 2, respectively, were 6.04 minutes, 4.00
minutes, and 5 minutes. As discussed previously, however, a significant bias for a segment of the WTC 2
population may exist: occupants above floor 78 whose evacuation delay time exceeded 16 minutes did not
successfully evacuate WTC 2 on September 11, 2001. NIST has estimated that approximately 75 percent
of the occupants above the 78th floor at 8:46:30 a.m. were able to evacuate on September 11, 2001,
therefore this bias accounts for less than 10 percent of the overall WTC 2 population.
The Structural Equations for the Model
The theoretical model presented in Figure 1 was represented by the following structural equations:
X 4 = p 4 iXi + p 42 X 2 + P43X3 + e 4 (Eq.C-1)
X 5 = p 51 X! + p 52 X 2 + P53X3 + P54X4 + e 5 (Eq. C-2)
X 6 = p 61 X, + p 62 X 2 + p 63 X 3 + p 64 X 4 + p 65 X 5 + e 6 (Eq. C-3)
X 7 = foX, + p 72 X 2 + p 73 X 3 + p 74 X 4 + p 75 X 5 + p 76 X 6 + e 7 (Eq. C-4)
These equations cast perceived risk (X 4 ) as a direct linear function of environmental cues (Xi), floor (X 2 ),
and obtained information (X 3 ). Seeking information (X 5 ) is a direct linear function of environmental cues
(Xx), floor (X 2 ), obtained information (X 3 ), and perceived risk (X 4 ). Pre-evacuation actions (X 6 ) is a direct
linear function of environmental cues (Xi), floor (X 2 ), obtained information (X 3 ), perceived risk (X 4 ), and
seeking information (X 5 ). Finally, evacuation delay (X 7 ) was cast as a direct linear function of
environmental cues (Xi), floor (X 2 ), obtained information (X 3 ), perceived risk (X 4 ), seeking information
(X 5 ), and pre-evacuation actions (X 6 ).
Estimation of the Model and Assessing for Regression Assumptions
The model was estimated on the data from both WTC 1 and WTC 2. The estimated model parameters
included path coefficients (P), explained variance for each equation (R 2 ), and other estimates. These are
presented in Table C-l and Table C-2, respectively, for WTC 1 and WTC 2.
The models were assessed for specification error, multicolinearity, nonlinearity, and heteroscedasticity in
order to determine if basic regression assumptions could be reasonably made, and if the estimated model
parameters were unbiased.
Specification error was not determined to be a problem. The model included only major variables of
import suggested by evacuation theory and excluded variables shown in our preliminary analyses as non-
predictive in the data sets being analyzed.
Multicolinearity can bias model estimates because it can increase the standard errors of estimated
regression coefficients. The models for both WTC 1 and WTC 2 were assessed for multicolinearity in two
ways. First multicolinearity does not impose a problem unless it is nearly perfect. The zero-order
correlation matrices for both models (Table C-3 and Table C-4) were inspected to determine if any
NIST NCSTAR 1-7, WTC Investigation 221
Appendix C
Table C-l
Estimated
parameter of the model 1
for WTC 1.
Variables"
Path
Equation
Endogenous
Exogenous
Coefficient
Estimate
a
a
R 2
X 4
Xi
P41
.38
.00
.00
.55
X 2
P42
.37
.00
x 3
043
.14
.00
x 5
Xi
Psi
.21
.00
.00
.25
x 2
P52
.21
.00
x 3
P53
.01
N/S
x 4
P54
.15
.02
x 6
Xi
P61
.27
.00
.00
.68
x 2
P62
.41
.00
x 3
P«
.05
.06
x 4
P64
.08
.04
x 5
P65
.20
.00
x 7
Xi
Pvi
.29
.00
.00
.49
x 2
P72
-.17
.00
x 3
P 73
.20
.00
x 4
P74
-.02
N/S
x 5
P 75
.10
.01
x 6
P76
.47
.00
a. Where X] = environmental cues, X 2 = floor, X 3 = obtained information, X 4 = perceived risk, X 5 = sought information,
X 6 = pre-evacuation actions, and X 7 = delay in evacuation initiation.
222
NIST NCSTAR 1-7, WTC Investigation
Causal Modeling
Table C-2.
Estimated parameters of the model for WTC 2.
Variables"
Path
Equation
Endogenous
Exogenous
Coefficient
Estimate
a
a
R 2
X 4
Xi
041
.30
.00
.00
.59
x 2
P42
.49
.00
x 3
P«
.12
.00
x 5
Xi
Psi
.25
.00
.00
.25
x 2
P 5 2
.11
N/S
x 3
P53
.07
N/S
x 4
P 5 4
.17
.02
x 6
Xi
p61
.20
.00
.00
.69
x 2
P62
.36
.00
x 3
P«
.07
.04
x 4
P64
.23
.00
x 5
P65
.17
.00
x 7
Xi
Pvi
.13
.01
.00
.56
x 2
P72
-.19
.00
x 3
P 73
.23
.00
x 4
P 7 4
.05
N/S
x 5
P75
.11
.01
x 6
P76
.51
.00
a. Where X] = environmental cues, X 2 = floor, X 3 = obtained information, X 4 = perceived risk, X 5 = sought information,
X 6 = pre-evacuation actions, and X 7 = delay in evacuation initiation.
NISTNCSTAR 1-7, WTC Investigation
223
Appendix C
Table C-3. Zero-order correlation matrix for WTC 1.
Xi
x 2
x 3
x 4
x 5
x 6
x 7
x,
1
.68
.25
.67
.45
.71
.59
x 2
-
1
.26
67
.45
.76
.46
x 3
-
-
1
.34
.17
.29
.37
x 4
-
-
-
1
.43
.64
.47
x 5
-
-
-
-
1
.55
.44
x 6
-
-
-
-
-
1
.64
x 7
-
-
-
-
-
-
1
a. Where X[ = environmental cues, X 2 = floor, X 3 = obtained information, X 4 = perceived risk, X 5 = sought information,
X 6 = pre-evacuation actions, and X 7 = delay in evacuation initiation.
Table C-4. Zero-order correlation matrix for WTC 2. a
Xi
x 2
x 3
x 4
x 5
x 6
x 7
x t
1
.60
.43
.64
.46
.67
.56
x 2
-
1
.44
.72
.41
.75
.48
x 3
-
-
1
.46
.30
.47
.55
x 4
-
-
-
1
.44
.72
.55
x 5
-
-
-
-
1
53
.47
x 6
-
-
-
-
-
1
.68
x 7
-
-
-
-
-
-
1
a. Where X] = environmental cues, X 2 = floor, X 3 = obtained information, X 4 = perceived risk, X 5 = sought information,
X 6 = pre-evacuation actions, and X 7 = delay in evacuation initiation.
correlations between the regressors where around .80 or higher — this is a typical cutoff value below
which multicolinearity does not seriously bias model estimates. An inspection of the zero-order
correlations in Table C-3 and Table C-4 led to the conclusion that coefficients were not sufficiently high
for multicolinearity to be a problem. Second, multicolinearity was assessed by regressing each exogenous
variable in each equation on all other exogenous variables in that equation, and the explained variances
for these regressions were inspected to see if any approached 1.00, which would indicate a biasing level
of multicolinearity. This assessment also led to the conclusion that multicolinearity was not a source of
224
NIST NCSTAR 1-7, WTC Investigation
Causal Modeling
bias in the estimated parameters of the model in either the data set for WTC 1 or the data for WTC 2.
Therefore, it was concluded that the regressors in both models were orthogonal.
The models were then assessed to determine if the assumption of linearity could be met. Exogenous
variables in each equation were transformed to alternative nonlinear forms, for example, the natural
logarithm of X, the square of X, the reciprocal of X, and the square-root of X. These transformed
variables were then correlated with each of the pre-determining and endogenous variables in both models.
None of the correlations involving the transformed exogenous variables increased substantially beyond
the linear correlations presented in Table C-3 and Table C-4. A visual inspection of scatter plots also
confirmed the conclusion that relationships were linear.
The assumption of homoscedasticity was also assessed by visual inspection of regression residuals in
scatter plots for each relationship in both models; and it was concluded that this assumption was met.
The observed means, standard deviations, and ranges for all of the variable included in the models for
both towers are presented in Table C-5.
Table C-5. Observed means, standard deviations, and ranges for variable
measurements in WTC 1 and WTC 2.
Variable 3
WTC1
WTC 2
Mean
SD
Range
Mean
SD
Range
x,
.84
.73
2.00
.68
.70
2.00
x 2
41.87
25.71
90.00
48.99
30.56
104.00
x 3
.11
.32
1.00
.21
.41
1.00
x 4
.63
.48
1.00
.67
.47
1.00
x 5
.28
.44
1.00
.29
.45
1.00
x 6
2.08
1.50
7.00
2.37
1.46
8.00
x 7
5.61
8.34
79.00
6.04
8.06
74.00
a. Where X] = environmental cues, X 2 = floor, X 3 = obtained information, X 4 = perceived risk, X 5 = sought
information, X 6 = pre-evacuation actions, and X 7 = delay in evacuation initiation.
Judging the Success of the Models
The estimated parameters of the models for WTC 1 and WTC 2 (see Tables 1 and 2) revealed that the
model had a very high degree of success in explaining evacuation initiation delay, pre-evacuation actions,
seeking information, and perceived risk in both towers. The adjusted explained variance (R 2 ) for
perceived risk was 55 and 60 percent in WTC 1 and WTC 2, respectively. The adjusted explained
variance for sought information was 25 percent in both towers. Respectively, the adjusted R 2 for pre-
evacuation actions was 68 and 69 percent for WTC 1 and WTC 2, respectively. Finally, the adjusted R 2
for delay in evacuation for WTC 1 and WTC 2 was, respectively, 49 and 56 percent. These are
extraordinarily high levels of adjusted explained variance to observe in a study of human evacuation and
these R 2 s, thereby, establish the strong predictive power of the models for both towers. All of the
equations in the models for both towers were statistically significant at the .001 level or higher.
NISTNCSTAR 1-7, WTC Investigation
225
Appendix C
Direct Effects
With a few exceptions, similar findings emerged in both towers. This lends validity to the conclusions
that can be drawn from the analyses. The findings are presented below. First, we consider each of the four
equations in both models for both towers (see Figure 1), and then the models are interpreted as a whole so
that the most significant paths of influence for each tower can be distinguished.
Predicting perceived risk. The findings that emerged regarding predicting the risk that people perceived
were virtually identical across the two towers. The R 2 for perceived risk was 55 percent in WTC 1 and
60 percent in WTC 2 (see Tables 1 and 2). In WTC 1, both environmental cues and floor had strong and
similar impacts on predicting perceived risk, respectively, p 4) and p 42 were .38 and .37, while obtained
information had a weaker but statistically significant impact, p 43 was . 14. In WTC 2, once again, both
environmental cues and floor had strong impacts on predicting perceived risk, respectively p 4 i and p 4 2
were .30 and .49, and, once again, obtained information had a weaker but statistically significant impact,
P« was .12. These findings suggest that the risk that people perceived before they began their evacuation
increased largely as a function of floor height and being exposed to environmental cues. It is most likely
that floor had this effect due to the perceived increased time needed to evacuate as a result of being higher
in the towers. Environmental cues likely had this effect since seeing, hearing, feeling, and so on physical
cues that indicate danger make discounting danger — most people's natural inclination—harder to achieve.
Obtained information likely increased perceived risk because people learned more about the seriousness
of the event through the information they obtained. Clearly, however, information had a lesser impact on
risk perception than did the two more salient variables of experiencing environmental cues and floor
height. The only difference in findings between the towers was that, in WTC 2, floor height was by far
the strongest predictor of perceived risk.
Predicting seeking information. Once again, the findings that emerged for predicting seeking
information were, almost, identical across both towers. Explained variance (R 2 ) for seeking information
was 25 percent in both towers. In WTC 1, environmental cues and floor both had the strongest and
identical impacts on seeking information, p 5[ and p 5 2 were both .21; obtained information had no impact
on seeking information, p 53 was not statistically significant; and perceived risk had a slight impact on
seeking information, p 54 was .15. In WTC 2, environmental cues had the strongest impact on seeking
information, p 5[ was .25; the impacts of floor and obtained information, p 52 and p 53 , were not significant;
and perceived risk had a slight impact on seeking information, p 54 was .17. Seeking information in times
of rapid onset emergencies is a typical human response since people need to interpret and make sense out
of an event before they act on it. The finding that environmental cues were the strongest predictor of
seeking additional information is consistent with this theoretical finding about "milling" from past
research. Obtained information had no impact on seeking information in either tower. This was likely
because information to make sense out of the event had already been obtained. Perceived risk had a
similar effect on seeking information — albeit lesser of an effect than environmental cues — in both towers,
likely because it increased the urgency people had to interpret the situation. Interestingly, floor height (or
distance from the exit) had a significant effect on seeking information in WTC 1, but not in WTC 2.
Evacuation theory would predict that this effect would be present, it was in the tower that was struck first,
and it was not present in the tower stuck second likely because the event began for the occupants of
WTC 2 long before the second plane struck their tower.
226 NIST NCSTAR 1-7, WTC Investigation
Causal Modeling
Predicting pre-evacuation actions. The explained variance (R 2 ) for taking pre-evacuation actions in
WTC 1 was 68 percent, and it was 69 percent in WTC 2 (See Tables 1 and 2). In both towers, the
strongest predictor of taking pre-evacuation actions was floor, p 6 2 in WTC 1 was .41 and it was .36 in
WTC 2. Environmental cues was also predictive of pre-evacuation actions, p 6 i, respectively, in WTC 1
and WTC 2 was .27 and .20, respectively. Once again, observing clues that one is at risk and being high
in the building with a longer path to safety emerged as strong predictors, in this case of taking actions to
ready to leave. Obtaining information had virtually no impact in either tower; p 63 was .05 in WTC 1 and
.07 in WTC 2. Seeking information impacted pre-evacuation action, p 65 was .20 in WTC 1 and it was
.17 in WTC 2; likely because the information obtained supported the need to evacuate and, hence, related
to getting ready to leave. Finally, the impact of perceived risk on taking pre-evacuation actions (p 6 4) was
.23 in WTC 2, but it was weaker in WTC 1 where is was .08.
Predicting delay in evacuation. Explained variance (R 2 ) in evacuation delay was 49 percent in WTC 1
and 56 percent in WTC 2. The impacts of environmental cues (p 7i ), floor (p 72 ), obtained information (p 73 ),
perceived risk (P74), sought information (p 75 ), and pre-evacuation action (p 76 ) on delay in evacuation
initiation, respectively, were .29, -.17, .20, -0.2, .10, and .47 for WTC 1; and for WTC 2 they were
.13, -.19, .23, .05, .11, and .51. The greatest predictor of evacuation delay in both towers was taking pre-
evacuation actions. Obviously, doing anything before initiating evacuation — including things to ready to
leave — delayed departure. Setting this factor aside, some clear differences emerged between the two
towers in terms of the relative impacts of the remaining variables in the model. Perceived risk (P 74 ) had
no direct effect on evacuation initiation delay. This finding is consistent with general evacuation theory
where perceived risk's impact on actual behavior is indirect through other factors. The three factors with
the strongest direct effects on evacuation delay were the same in both towers. These were environmental
cues, floor, and obtained information. In both towers, floor's effect was negative, that is, the more floors
one was from the exit, the quicker people were to initiate their evacuation. Environmental cues and
information that was received passively both increased delay in the initiation of evacuation. Finally
seeking additional information had a minimal impact on evacuation delay.
Paths of Greatest Influence and Conclusions
Although each of the aforementioned findings are interesting in and of themselves, perhaps the most
important findings that we can offer are those that emerge when all of the individual findings offered
above are brought together and viewed at the same time in the context of the entire model.
Bias. As discussed previously, any conclusions about evacuation initiation delay time in WTC 2 should
consider the impact of disproportional decedent location, particularly as a source of the disproportionality
may be highly correlated to the variable of interest, evacuation delay. In other words, those who
exhibited long delay times in one region of the building were unable to be interviewed, thus artificially
shortening the average delay time for one-third of the building. In the causal modeling, this would affect
the relationship between 'floor' and 'delay initiating evacuation,' likely tending towards zero a slightly
negative estimate (-0.19) of the beta value between the two variables. As floor was not a primary path
which directly predicted evacuation initiation delay in WTC 2, the impact of this bias was considered
secondary. The effect of this bias as it worked through other variables was not considered.
NIST NCSTAR 1-7, WTC Investigation 227
Appendix C
WTC 1. Although there were other factors that had some lesser impacts on influencing what people did,
the paths of causal influence that defined the main process that led to delay in the evacuation of WTC 1
on September 11, 2001, follows.
Environmental cues (information from the physical environment that something was terribly wrong) and
floor (increased distance to safety) caused people to set out to find additional information, most likely
information about what was going on and what they should do. Next, the act of seeking additional
information, that is, "milling" about to make sense out of the situation, led people to take actions to
prepare to evacuate. Finally, taking those actions to prepare to evacuate delayed the initiation of actually
evacuating.
In addition to this four step causal process, environmental cues and floor also had indirect impacts on
evacuation delay as follows. Both factors increased the odds of seeking information and both factors
increased the chances that people would take pre- evacuation actions prior to evacuating. Both factors also
had direct impacts on actual evacuation delay. Environmental cues increased delay while floor decreased
delay.
WTC 2. Although there were other factors that had lesser influence on what people did, the paths of
causal influence that defined the main process that led people to delay in the evacuation of WTC 2 on
September 11, 2001 were identical to WTC 1 with one decided difference.
Environmental cues (information from the physical environment that something was terribly wrong) and
floor (increased distance to safety) predicted perceived risk. Environmental cues, floor, and perceived
risk caused people to set out to find additional information. Next, the act of seeking additional
information, that is "milling" about to make sense out of the situation, and perceived risk both led people
to take actions to prepare to evacuate. Finally, taking those actions to prepare to evacuate delayed the
initiation of actually evacuating.
In addition to this five step causal process, environmental cues and floor also had indirect impacts on
evacuation delay. Both factors increased the odds of seeking information and both factors increased the
chances that people would take pre-evacuation actions prior to evacuating. Both factors also had direct
impacts on actual evacuation delay. Environmental cues increased delay, while floor decreased delay.
C.2 Predicting Normalized Stairwell Evacuation Time In WTC 1 On
September 11, 2001
This project analyzed the factors and social processes that influenced the normalized stairwell evacuation
time per story of stairs for the people who evacuated out of WTC 1 on September 11, 2001. WTC 2 was
excluded from this analysis because evacuees used stairs, elevators, and/or a combination of both for their
evacuation. Evacuation time was defined as the average number of seconds per story of stairs that it took
people from the time they entered a stairwell until they completed their evacuation out of the building.
For a discussion of the data collection method, see Section 10.2.1 of this report.
228 NIST NCSTAR 1-7, WTC Investigation
Causal Modeling
Pre-Modeling Analysis
The model estimated to predict normalized stairwell evacuation time was specified as follows:
First, we accessed the record of research on human evacuation during high rises fires. A group of
individuals selected for the expertise on human evacuation during high rise fires was assembled. The
individual experts developed the final list of variables, and the relationships between them, to use to
specify the model used in this analysis.
Second, preliminary statistical analyses were performed to determine which variables thought to be
important predictors of normalized stairwell evacuation time by the team of experts had and did not have
predictive salience in WTC 1. Those that had no significant impact on normalized stairwell evacuation
time were excluded from further consideration. The factors that had salient impacts on normalized
stairwell evacuation time were carried forward into the model used in this analysis.
The Model that was Estimated
The model used to predict normalized stairwell evacuation time in WTC 1 of the World Trade Center on
September 11, 2001, is diagrammatically illustrated in Figure C-2. This model can be described as
follows: (1) normalized stairwell evacuation time is a direct consequence of floor, evacuation decision
delay, environmental cues, emergency responders, crowding, and evacuation interruption; (2) evacuation
interruption is a direct consequence of floor, evacuation decision delay, environmental cues, emergency
responders, and crowding; (3) crowding is a direct consequence of floor, evacuation decision delay,
environmental cues, and emergency responders; (4) emergency responders is a direct consequence of
floor, evacuation decision delay, and environmental cues; (5) environmental cues is a direct consequence
of floor and evacuation decision delay; and, finally, (6) evacuation decision delay is a direct consequence
of floor. This model is parsimonious, consistent with the input received from high rise fire evacuation
experts, and it well-represented the positive findings of our preliminary analyses of the many variables
that could have impacted normalized stairwell evacuation time.
Measurement
Since multiple regression analysis requires that all responses be analyzed in numerical form, answers had
to be 'coded' or converted to numbers. By convention, = 'no', and 1 = 'yes.' Further, 2 is always a
greater quantity of whatever the measurement is evaluating than 1.
Floor (Xi) was measured by asking respondents which floor they were on when the event started.
Responses were coded from negative seven (basement floors) to 105. Negative floor values were
transposed into positive ones since this measure sought to determine how many floors people were from
their building's floor of exit. Missing data was coded to the mean.
Evacuation decision delay (X 2 ) was measured by asking respondents how minutes passed from the time
when they first became aware that something was wrong until they actually began their evacuation. Their
responses were coded as an interval scale of minutes.
NIST NCSTAR 1-7, WTC Investigation 229
Appendix C
e 4
Evacuation
Decision
Emergency
Responders
Evacuation
Interruption
Normal
Stairwell
Evacuation
Time
Environmental
Cues
Xj
Crowding
x s
Figure C-2. Model of normalized stairwell evacuation time.
Environmental cues (X 3 ) was measured by asking respondents about the number and type of
environmental cues (severe signs of danger, and non-severe signs of danger) that they saw prior to
initiating their evacuation. Answers to these questions were coded as dummy variables (0 or 1) and then
added to form an environmental cues scale that could vary between and 2.
Emergency responders (X 4 ) was measured by asking respondents whether items from a list made their
evacuation more difficult. Affirmative answers to the "firefighters/police" choice were coded as 1, and
"no" answers and missing data were coded as 0.
Crowding (X 5 ) was measured by asking respondents whether items from a list made their evacuation
more difficult. "Yes" answers to "crowded stairwells" were coded as 1, and "no" answers and missing
data were coded as 0.
Evacuation interruption (X 6 ) was measured by asking respondents whether they turned back at any time
during their evacuation. "Yes" answers were coded as 1 , and "no" answers and missing data were coded
asO.
Finally, normalized stairwell evacuation time (X 7 ) was measured by asking respondents how much time
passed between the moment they first began their evacuation until they exited the tower. Answers were
coded on an interval scale; "don't know" and "refused to answer" responses were coded to the mean.
These raw numbers were then normalized. The hydraulic model of people movement dominates the
average evacuation time per floor, so it had to be removed from normalized stairwell evacuation time so
as to fully explain the fluctuations (denoted by a prime) in reported evacuation time. This was
230
NIST NCSTAR 1-7, WTC Investigation
Causal Modeling
accomplished by a least squares analysis of reported evacuation time against floor for each tower, with
statistical outliers removed. The final WTC 1 estimate was:
AETF'j = RETj - .056; (Eq. C-5)
where:
AETF' = Average fluctuation in evacuation time per floor for occupant j,
i = Floor evacuation began from for occupant j, and
RET = Reported evacuation time for occupant j.
The Structural Equations for the Model
The theoretical model presented in Figure 1 was represented by the following structural equations.
X 2 = p 21 X! + e 2 (Eq. C-6)
X 3 = p 3 iXi + p 32 X 2 + e 3 (Eq. C-7)
X 4 = p 4 iXj + p 42 X 2 + p 43 X 3 + e 4 (Eq. C-8)
X 5 = p 51 X! + p 52 X 2 + P53X3 + P54X4 + e 5 (Eq. C-9)
X 6 = p 61 Xi + p 62 X 2 + p 63 X 3 + p 64 X 4 + p 65 X 5 + e 6 (Eq. C-10)
X 7 = p 71 X! + p 72 X 2 + p 73 X 3 + p 74 X 4 + p 75 X 5 + p 76 X 6 + e 7 (Eq. C-l 1)
These equations cast evacuation decision delay (X 2 ) as a direct linear function of floor (Xi).
Environmental cues (X 3 ) was set as a direct linear function of floor (Xi) and evacuation decision delay
(X 2 ). Emergency responders was seen as a direct linear function of floor (Xj), evacuation decision delay
(X 2 ), and environmental cues (X 3 ). Crowding was cast as a direct linear function of floor (Xi), evacuation
decision delay (X 2 ), environmental cues (X 3 ), and emergency responders (X 4 ). Evacuation interruption
(X 6 ) was cast as a direct linear function of floor (Xi), evacuation decision delay (X 2 ), environmental cues
(X 3 ), emergency responders (X 4 ), and crowding (X 6 ). Finally, normalized stairwell evacuation time was
cast as a direct linear function of floor (Xi), evacuation decision delay (X 2 ), environmental cues (X 3 ),
emergency responders (X 4 ), crowding (X 5 ), and evacuation interruption (X 6 ).
Estimation of the Model and Assessing for Regression Assumptions
The model was estimated on the data from WTC 1. The estimated model parameters included path
coefficients (betas), explained variance for each equation, and other estimates. These are presented in
Table C-6. The zero-order correlations between each of the variables are presented in Table C-7.
NIST NCSTAR 1-7, WTC Investigation 23 1
Appendix C
Table C-6.
Estimated parameters of the model for WTC 1.
Variables"
Path
Equation
Endogenous
Exogenous
C oeff icient
Estimate
a
a
R 2
X 2
Xi
B 2[
.58
.00
.00
.34
x 3
X,
B 3]
.78
.00
.00
.79
x 2
B 32
.17
.00
x 4
Xi
041
.23
.00
.00
.57
x 2
042
-.01
N/S
x 3
P«
.56
.00
x 5
Xi
Psi
.18
.00
.00
.72
x 2
P 5 2
.03
N/S
x 3
P53
.37
.00
x 4
P54
.35
.02
x 6
x,
Pel
.12
N/S
.00
.11
x 2
P62
.04
N/S
x 3
P 63
.15
N/S
x 4
P64
.01
N/S
x 5
P65
.07
N/S
x 7
x,
P71
.03
N/S
.00
.44
x 2
P72
-.05
N/S
x 3
P73
.46
.00
x 4
P74
.09
N/S
x 5
P 75
.08
N/S
x 6
P76
.18
.00
a. Where X] = floor, X 2 = evacuation delay, X 3 = environmental cues, X 4 = emergency responders, X 5 = crowding, X 6 ;
evacuation interruption, and X 7 ^normalized stairwell evacuation time.
232
NIST NCSTAR 1-7, WTC Investigation
Causal Modeling
Table C-7. Zero-order correlation matrix for WTC 1.
x,
X 2
x 3
x 4
x 5
x 6
x 7
X,
1
.58
.88
.71
.77
.84
58
x 2
-
1
.63
.47
.53
.25
.34
x 3
-
-
1
.75
.81
.35
.58
x 4
-
-
-
1
.77
.28
.54
x 5
-
-
-
-
1
.31
.57
x 6
-
-
-
-
-
1
.39
x 7
-
-
-
-
-
-
1
a. Where X] = floor, X 2 = evacuation delay, X 3 = environmental cues, X 4 = emergency responders, X 5 = crowding, X 6 =
evacuation interruption, and X 7 ^normalized stairwell evacuation time.
The model was assessed for specification error, nonlinearity, and heteroscadasticity in order to determine
if basic regression assumptions could be reasonably made, and if the estimated model parameters were
unbiased. The multicolinearity assumption did not have to be assessed since the model contained only one
exogenous variable.
Specification error was determined not to be a problem. The model included only variables of import
suggested by individual experts, and excluded variables shown in our preliminary analyses as non-
predictive in the data sets being analyzed.
The model was then assessed to determine if the assumption of linearity could be met. Exogenous
variables in each equation were transformed to alternative nonlinear forms, for example, the natural
logarithm of X, the square of X, the reciprocal of X, and the square-root of X. These transformed
variables were then correlated with each of the pre-determining and endogenous variables in the model.
None of the correlations involving the transformed exogenous variables increased substantially beyond
the linear correlations presented in Table C-7. A visual inspection of scatter plots also confirmed the
conclusion that relationships were linear.
The assumption of homoscedasticity was assessed by visual inspection of regression residuals in scatter
plots for each relationship in both models, and it was concluded that this assumption was met.
The observed means for each of the variables contained in the model were: 47.2 for floor, 5.36 for
evacuation delay, 1.73 for environmental cues, 0.60 for emergency responders, 0.73 for crowding, 0.12
for evacuation interruption, and 13.01 for normalized stairwell evacuation time.
NISTNCSTAR 1-7, WTC Investigation
233
Appendix C
Judging the Success of the Models
The estimated parameters of the model for WTC 1 (Table C-6) revealed that the model had a very high
degree of success in explaining normalized stairwell evacuation time. The adjusted explained variance
(R 2 ) for normalized stairwell evacuation time was 44 percent, 1 1 percent for evacuation interruption, 72
percent for crowding, 57 percent for emergency responders, 79 percent for environmental cues, and 34
percent for evacuation initiation delay. With the exception of evacuation interruption, these are
extraordinarily high levels of adjusted explained variance to observe in a study of human evacuation.
These R 2 's, thereby, establish the strong predictive power of the model. All of the equations in the model
were statistically significant at the .001 level or better.
Direct Effects in the Model
We first consider each of the six equations in the model (see Figure 1), and then the model is interpreted
as a whole so that the most significant paths of influence can be distinguished.
Predicting evacuation initiation delay . The findings that emerged regarding predicting delay in the
initiation of evacuation from floor were that the R 2 was 34 percent with a value of .58 for p 2 i significant at
the .001 level. This relationship was elaborated earlier in this appendix.
Predicting environmental cues . Explained variance (R 2 ) for observing environmental cues was
79 percent, and the equation was statistically significant at the .001 level. Floor had a very strong direct
impact on observing environmental cues; p 3i was .78 and it was significant at the .001 level. The effect of
delay in the initiation of evacuation (p 32 ) was .17, and it was also statistically significant at the .001 level.
It appears that the longer a person took to begin their evacuation, the more the physical impacts of the
event grew and the more likely people were to experience them.
Predicting emergency responders . The explained variance (R 2 ) for the third equation in the model
predicting encountering emergency responders was 57 percent, and the equation was statistically
significant at the .001 level. The relative effects of floor, delay in beginning evacuation, and
environmental cues, respectively, were as follows: p 4i was .23 statistically significant at the .001 level, P42
was -.01 and was not statically significant, and p 43 was .56 statistically significant at the .001 level. Floor
height and experiencing environmental cues both predicted encountering emergency responders. This
makes sense when one considers that emergency responders would be most likely to go to areas
experiencing the impacts that would also yield environmental cues, and the higher one was in the tower,
the more stairwells one had to traverse and the greater the odds of encountering emergency responders.
Predicting crowding . The explained variance (R 2 ) for the fourth equation in the model that predicted
perceived crowding on the evacuation stairwells was 72 percent. The four variables in the equation had
the following effects on perceived crowding: floor (p 5 i) was .18, significant at the .001 level; evacuation
delay (p 52 ) was .03, and it was not statistically significant; environmental cues (p 53 ) was .37, significant at
the .001 level; and encountering emergency responders (p 54 ) was .35, significant at the .001 level.
Perceived crowding largely increased as a result of environmental cues and encountering emergency
responders.
Predicting evacuation interruption . The explained variance (R 2 ) for interrupting evacuation was only
1 1 percent. Even though this equation was statically significant at the .001 level, none of the 5 predictor
234 NIST NCSTAR 1 - 7, WTC Investigation
Causal Modeling
variables in the equation had a statistically significant impact on evacuation interruption. The beta
weights for these variables (floor, evacuation delay, environmental cues, emergency responders, and
crowding) were, respectively (p 6 i through p 65 ) .12, .04, .15, .01, and .07. At best, we can interpret these
findings to mean that there was a slight but not statistically significant tendency for people to interrupt
their evacuation if they had more rather than fewer floors to traverse to safety, and if they encountered
environmental cues (perhaps obstacles) in the process of evacuation.
Predicting normalized stairwell evacuation time . Explained variance (R 2 ) in predicting normalized
stairwell evacuation time was 44 percent, and the equation was statistically significant at the .001 level.
The impacts of floor (p 7 i), evacuation delay (P72), environmental cues (P73), emergency responders (P74),
crowding (p 75 ), and evacuation interruption (p 76 ), on normalized stairwell evacuation time, respectively,
were .03, -.05, .46, .09, .08, and .18. Of these, only two factors in the equation were statistically
significant both at the .001 level. These were environmental cues (p 73 was .46) and evacuation interruption
(P 76 was .18). Clearly, the single factor that had the biggest impact on increasing the amount of time
people spent, on average, per stairwell was environmental cues. The only other factor that had a
significant impact was interrupting evacuation, obviously, because stopping egress would increase the
amount of time needed to complete evacuation.
Conclusions
Although each of the aforementioned findings are interesting in and of their own right, the most important
findings that we can offer are those that emerged when all of the above findings are brought together and
viewed at the same time in the context of the model as a whole (see Figure 1). What doing so revealed
was that, based on this analysis, the main process that led to increased normalized stairwell evacuation
time in the evacuation of World Trade Center WTC 1 on September 11, 2001, was straightforward and
clear. It can be described as follows.
Floor (increased distance to safety) substantially increased the odds that people would encounter
environmental cues. Floor also increased delay in starting evacuation, which, in turn, also increased the
chances that people would encounter environmental cues. But it was encountering environmental cues
(which likely blocked egress) that had a large and direct effect on increasing the amount of time that
people spent, on average, to traverse their evacuation stairwell. In addition to this multi-step process with
environmental cues as the key predicting variable, interrupting the process of evacuation for any reason
also increased the amount of time, on average, that people used to descend their evacuation stairwell.
NISTNCSTAR 1-7, WTC Investigation 235
Appendix C
This page intentionally left blank.
236 NIST NCSTAR 1-7, WTC Investigation
Appendix D
Egress Modeling
The purpose of modeling evacuation from the World Trade Center (WTC) towers was to obtain
evacuation times for a variety of scenarios which provide context to understanding the actual evacuation
on September 11, 2001. The following six scenarios were selected and modeled:
Phased evacuation
• Scenario 1 : Occupants in an emergency situation have to travel 3 floors below the fire floor, floor
48
Total evacuation
Scenario 2: Full capacity tower without visitors or damage
Scenario 3: Full capacity tower including visitors without damage
Scenario 4: Sept. 1 1th capacity tower without damage
Scenario 5: Full capacity tower with plane damage blocking floors 91 and above in WTC 1
Scenario 6: Full capacity with plane damage blocking floors 78 and above in WTC 2 after
16 minutes, including estimates of elevator usage
D.l METHOD FOR THE SIMULATING PHASED EVACUATION (SCENARIO 1)
The purpose of simulating a phased evacuation was to understand how the evacuation strategy would
have worked during an emergency at the WTC towers. For a phased evacuation, the occupants of the
emergency floor, the occupants on the floor above, and the occupants on the floor below were to evacuate
to three floors below the emergency floor. On floors 47 through 49, 200 occupants were placed on each
floor. These 600 occupants traveled to the 45th floor where they would be considered "safe." Floor 48,
the emergency floor, was a transfer floor; therefore, occupants from floor 49 were faced with moving
through the horizontal corridor on the 48th floor. Figure D-l shows a three-dimensional view of the
45th - 50th floors, as entered into the buildingEXODUS model. The occupants were placed on
floors 47 - 49 and the three stairwells connected all 6 floors together.
NIST NCSTAR 1-7, WTC Investigation 237
Appendix D
Figure D-l. 3-D view of a phased evacuation showing floors 45 - 50 (Exodus).
Figure D-2 shows occupants traveling through one of the transfer corridors on floor 48. Occupants are
traveling on the stairs that connects floor 49 and 48, walking through the transfer corridor and eventually
onto the stair connecting floor 48 to 47.
Figure D-2. Transfer corridor on floor 48 (Exodus).
D.2
METHOD FOR SIMULATING TOTAL EVACUATION (SCENARIOS 2-6)
Total building evacuation represented an emergency in which all occupants of the building evacuated
simultaneously. Scenarios 2-6 were total evacuation simulations, which varied the total occupant
population inside a tower and the damage to the building at the time of evacuation.
D.2.1
Population Estimates
Scenarios 2-4 required population estimates. Scenario 2 required an estimate of the number of
occupants in the building, without the inclusion of visitors. Furniture and egress floor plans were studied
from a sampling of floors: in WTC 1 - floors 94-100 and in WTC 2 - floor 77. On each floor, the
238
NIST NCSTAR 1-7, WTC Investigation
Egress Modeling
number of offices, cubicles, and other workstations was counted. From these numbers, the usable floor
area (square footage) of the floors was divided by the corresponding number of persons to obtain an
occupant density (area per person). Each floor in the towers received an average density (14 m 2 /person),
and the number of occupants per floor changed according to the total usable area of each floor. This
resulted in a total of 19,800 persons to be simulated in Scenario 2.
For scenario 3, the full-capacity population with visitors, the number of occupants inside the towers was
determined by the same method as Scenario 2, adjusted for visitor spaces. Visitor spaces included
conference rooms, waiting areas, lunch tables, and library chairs. The same floor plans were used from
Scenario 2 population development to count these additional spaces. Then, as in the previous calculation,
the usable floor area of each floor in the towers was used to calculate the number of occupants, including
visitors in each tower. With these calculations, the occupant density was 1 1 m 2 /person when including
visitors. This resulted in an estimate of 25,500 people. The findings of each step of the estimation
method are summarized in Table D-l. Note, however, that in 2005, PANYNJ estimated that the
maximum population of WTC 1 or WTC 2 would not likely have exceeded 20, 000. 68
Table D-l. Floor and occupant modeling parameters.
No. of chairs
No. of chairs (with
Zone
Represented Floors
Approx. area (m 2 )
(w/o visitors) 3
visitors)*
1
9-26
27-34
35-40
2,723
2,826
2,922
200
260
2
43-47
2,679
190
240
3
49-54
55-56
57-74
2,974
2,935
3,007
210
270
4
77-81
2,860
200
260
5
83-95
103 - 104
96-102
105 - 106
3,188
3,309
3,245
3,245
230
300
a. Averaged over variations within each zone.
The population for Scenarios 4, 5, and 6 was based upon the National Institute of Standards and
Technology (NIST) estimate of the number of occupants present at WTC 1 and WTC 2 on
September 11, 2001, or approximately 8,800. The purpose of Scenario 4 was to obtain a lower bound
evacuation time for a September 11, 2001 population in the absence of constraints to evacuation or
elevator usage. The 8,800 estimate was used to 'calibrate' (refine the model output to match the gross
evacuation characteristics from the actual evacuation of WTC 1 and WTC 2) the model for Scenarios 5
and 6, where additional occupants were added in combination with building damage (Galea 2004).
For the total building evacuation scenarios, occupants were placed on all floors of the building, excluding
mechanical spaces. Floors 1 through 107 were modeled (floors 108-110 were not normally occupied).
During each simulation, the occupants left their floors and traveled into one of the three stairwells, in
1 Bhol, Saroj. PANYNJ (September 21, 2005). Email from S. Bhol to S. Sunder in response to NIST question.
NISTNCSTAR 1-7, WTC Investigation
239
Appendix D
which they remained throughout the entire evacuation. The stairwells included all transfer corridors and
emptied into the Mezzanine or the lobby area, depending upon the stairwell. Occupants reached safety as
soon as they exited the tower (either to the outside or into the Concourse mall area).
Finally, the value of 8,800 occupants for the number of occupants in a WTC tower on
September 11, 2001 was based upon the statistical projection derived from the telephone survey data
described previously in Table 4-1.
D.3 WTC Tower Geometry Development
Each Tower, without including the basement floors, contained a total of 1 10 floors. In order to ease the
difficulty of floor plan input in each evacuation model, 1 1 representative floor plans were created and
replicated throughout the building. The core layout and stair position did not change significantly in
between the transfer floors. Therefore, knowing the floor location of core and stair walls, the tower was
divided into zones, the boundaries of which were transfer floors. Either tower (since they were identical
in core layout and stair position) was divided into 5 zones, in addition to the four significant transfer
floors (floors 42, 48, 76, and 82). The 1 1 representative floors were the Concourse, Mezzanine, a
representative floor from each of the 5 zones, and the 4 unique transfer floors. Using representative floor
plans greatly simplified floor plan input into each evacuation model used, with little loss of accuracy.
Even though floor duplication was used, each floor had to be created individually to achieve an accurate
tower description within the evacuation model. At and around the mechanical spaces, the floor to ceiling
measurements increased, meaning that throughout the building, the number of stair risers between floors
ranged from 18 to 26 risers throughout the tower. Therefore, each floor was developed individually to
ensure it contained the correct number of stairs to connect to the floor above. Also, even though
occupants were not placed on these floors, mechanical floors were added to each model to include the sets
of stairs that led throughout the mechanical spaces.
Three additional stairwell movements were not modeled, as the stairwell translation was less than 5 m:
floors 26-41 (Stair A) and 66-68 (Stairs A and C). Note that the stair movement on floors 26-41 was
input into the buildingEXODUS evacuation model only, as stair positioning was more important in this
model, as discussed below.
Table D-2 shows the major floors plans contained in each zone. In Zones 1, 3, and 5, there were minor
changes to the core walls from floor to floor, shown in the additional floor plans. The table outlines
which floors these changes take place. However, due to the size of the building and the relative
insignificance of a core wall change to the evacuation (since most of an occupant's travel time was spent
in the stairwells), the floor plan of the majority of floors within a zone was used as the representative floor
in each evacuation model. Table D-2 shows that floor plan Zone 1-1 was used as a representative of
Zone 1, Zone 3-3 as a representative of Zone 3, and Zone 5-1 as representative of Zone 5.
240 NIST NCSTAR 1-7, WTC Investigation
Egress Modeling
Table D-2. WTC floor plan with core walls and stairwell locations.
Zone 1: Floors 3-41
no
o II □
Zone 1-1: Floors 9-26
There is a slight change in
Stair A at floor 25.
**Used for all of Zone 1
Zone 1-2: Floors 27-34
Variations are found in the upper
right quadrant throughout this
zone.
Zone 1-3: Floors 35-41
Variations are found in the upper
right quadrant throughout this
zone.
Transfer Floor 42
b
^^
Stairs A and C transfer
View of Stair A transfer
Zone 2: Floors 43-47
Escalator positions: There is an escalator on the right side of the upper
right quadrant between floors 44-45; Escalator on left side of upper left
quadrant between floors 43-45 - these will not be used for evacuation
purposes.
b
n
Transfer Floor 48
h
nr In
Stairs A and C transfer
View of Stair C Transfer
NISTNCSTAR 1-7, WTC Investigation
241
Appendix D
Zone 3: Floors 49-75
Zone 3-1: Floors 49 -54
Zone 3-3: Floors 55-56
Zone 3-3: Floors 57-75
1) Variations are found in the
upper left and right quadrants
throughout this zone.
2) There is a slight change in
Stairs A, C between floors 66-68.
*Used for all of Zone 3
Transfer Floor 76
C
Stairs A, B, and C transfer
View of Stair B Transfer
Zone 4: Floors 77-81
Q
"1
Transfer Floor 82
^L
i-j,"
tr-
Stairs A and C transfer
View of Stair C Transfer
242
NIST NCSTAR 1-7, WTC Investigation
Egress Modeling
Zone 5: Floors 83-107
uu
c
Zone 5-1: Floors 83-95
Variations are found in the lower
left quadrant throughout this
zone.
*Used for all of Zone 5
Zone 5-2: Floors 103-104, 107
Variations are found in the lower
left quadrant at floor 107.
— ^
u
rV
Zone 5-3: Floors 96-102, 105-
106
Variations are found in the lower
left quadrant throughout this
zone.
D.3.1 Single-Tenant and Multi-Tenant Floor Plans
The floor plans shown in Table D-2 (used in the evacuation modeling) were single-tenant (open) floor
plans, which contained only the outer walls, the core walls, and the stairwells. This section evaluates the
quality of the assumption that the subdivision of a floor space into multiple tenancies did not significantly
change the model results when compared to modeling the space as a single open floor plan. The
assumption was based upon the premise that in a building as large as WTC 1 or WTC 2, most of the
average occupant's evacuation time would be spent in the stairwell. Specific multi-tenant floor plans
were chosen to study, developed based upon the demising wall layouts found in the WTC Space Book
Plan (PANYNJ 2001c). The Simulex model was used for this analysis.
An open, single-tenant floor plan from each zone (1-5) was simulated by placing the appropriate number
of occupants on each floor, running the model, and recording the numbers of occupants using each stair.
The percentage of occupants who used each stairwell was calculated for each zone. Then, by factoring in
the number of floors in each zone (that the sample floor represents), the weighted percentages of occupant
use of each stair was calculated for the entire building. These percentages are shown in on the left hand
side, labeled as the single-tenant floor plan percentages.
Using the Space Book Plans, at least three representative multi-tenant floor plans were selected from each
zone. Each multi-tenant floor plan contained at least 3 corridors, and/or at least 1 different companies.
In each of the five zones, the movement of occupants within the multi-tenant floor plans was simulated,
and the percentages of stair use were averaged over each zone. The number of multi-tenant floors and
single-tenant plans per zone were calculated and assigned their appropriate average percentages. As
shown in Table D-3, the weighted percentages of occupant use of each stairwell were calculated for the
entire building.
NISTNCSTAR 1-7, WTC Investigation
243
Appendix D
Table D-3. Stair use percentages for single-tenant versus multi-tenant floor plans.
Single-tenant Floor Plans only
Single- and Multi-tenant Floor Plans
Stairwell C
Stairwell B
Stairwell A
Stairwell C
Stairwell B
Stairwell A
32.96 %
31.71 %
35.33 %
33.50%
32.40 %
34.10%
Number of people in each stair with 19,800 population
6526
6279
6995
6633
6415
6752
Difference (Gross)
-107
-136
243
Difference (Percent)
-0.5 %
-0.68 %
1.2%
In summary Table D-3 shows that there was approximately 1 percent difference between assuming all
floors to be single tenant (open) floor and modeling both single- and multi-tenant floors with respect to
the choice of stairwells. Therefore, representative single-tenant (open) floor plans were used in the
subsequent evacuation modeling.
D.4
EVACUATION MODELS
Three evacuation models were used. The Simulex model (IES 2001) was used to perform the single-
tenant and multi-tenant floor plan analysis and the phased evacuation scenario (Scenario 1). The
buildingEXODUS (Gwynne et al. 1998) model was used for all scenarios from the WTC towers
(Scenarios 1-6). Finally, the EXIT89 model (Fahy 1999) was used for the phased evacuation and the total
evacuation scenarios (without damage) from the WTC towers (Scenarios 1-4). This section provides brief
descriptions of each of the three models used in this project as well as the reason for picking these three
among all other available evacuation models.
Simulex
The Simulex model was developed by Thompson, from IES, Inc. in Scotland, and is a widely used model
in the field of fire safety. The model has been validated and focuses on the movement of occupants
throughout a structure (IES 2001), (IES 2000). The model allows the user to specify a distribution of
occupant types, associated with a certain body size and unimpeded movement speed, throughout the
building. The model includes important occupant movements, such as overtaking, side-stepping, back-
stepping, and others. Simulex predicts that occupant movement throughout building spaces will slow (in
speed) as the inter-person distance between occupants and / or obstacles decreases. The model has fixed
limitations on the number of floors and exits that the model can handle, however the developer is willing
to expand these to meet the needs of the users. Also, Simulex provides a 2-D visualization of the
evacuation throughout the building and allows the user to input CAD drawings of the building for ease in
building geometry development.
The input required for this model included the following:
• Floor plans in the form of CAD drawings
• Connections of floor levels by stairways or ramps (involving user input on stair width and length)
• Distance maps that can be created to block certain exits or paths from groups of occupants
244
NIST NCSTAR 1-7, WTC Investigation
Egress Modeling
• Occupant movement characteristics for each individual or a group of individuals with a
corresponding body size, initial horizontal speed, and percentage decrease of speed on stairs (the
user can use default values provided by Simulex or create his / her own)
• Occupant delay times to be assigned from a random, triangular, or normal distribution
building EXODUS
buildingEXODUS (Gwynne et al 1999; 1998; 2001) was developed by Galea at the University of
Greenwich, UK. buildingEXODUS is a well-documented and well-validated model, claimed to be
limited only by the storage capacity of the computer used to run the program. The model contains six
submodels that work together to form the evacuation simulation: occupant, movement, behavior, hazard,
geometry, and toxicity submodels. Any time occupants wish to occupy the same node, or a node that is
already occupied, a conflict resolution time is added to their total evacuation time. The method of conflict
resolution is how the model simulates congestion throughout the building. Individual abilities of the
occupants and conflict resolution are probabilistic. The behavioral model makes route choices depending
upon the environmental situation of the building and then passes this information onto the movement
model. Familiarity and awareness variables can be assigned to exits, and drive and patience variables are
assigned to the occupant. These variables are random and difficult for the user to specify for any
evacuation. buildingEXODUS provides the user with 2-D and 3-D visualization capabilities of the
building evacuation and allows for the input of CAD drawings of the building for ease in building
geometry development.
The input required for this model included the following:
•
•
•
Floor plans in the form of CAD drawings (other methods are available to provide building
information)
Connections of floor levels via stairways, with an option for including effective width inside the
stairwell
Distribution of occupant types throughout the building (user can specify information about each
individual or group such as speed, ability, drive, patience, etc.)
Option of including special nodes at certain points in the building to regulate occupant paths
through building and flow through the exits/doors
• Option of including boundary nodes used to predict slower movement near walls, obstacles
• A series of simulation options such as stair packing, occupant randomization throughout the
building, response time distribution, stair edge preference, and extreme behavior
EXIT89
EXIT89 (Fahy 1999), (Fahy 2001) is an evacuation model developed by Fahy at the National Fire
Protection Association (NFPA). The model has been tested on tall buildings in the past and has the
capability of modeling a large number of occupants in a building. The model relies on the density versus
NIST NCSTAR 1-7, WTC Investigation 245
Appendix D
speed data from Predtechenskii and Milinskii (Predtechenskii and Milinskii 1969) for different building
components, such as horizontal components, doorways, up stairs, and down stairs. It also uses
conditional movement, depending upon the presence and density of smoke in the evacuation path. All
occupants in the building are made up of the same body size and density versus speed distribution, both
specified by the user. EXIT89 uses a series of nodes and arcs instead of CAD drawings to represent any
type of structure. In this type of model, the floor plan is entered as a series of nodes (rooms, corridor,
stair sections, etc.) and arcs (distance between nodes). When using a node/arc model for a single-tenant
(open) floor plan such as those used for the WTC Towers, the user determines how to segment and link
the building space, and then should check to make sure that occupants are traveling in realistic patterns to
the exits. While EXIT89 does not provide visualization capability, the model does provide a particularly
detailed output that notes where each occupant is at each time period throughout the evacuation. EXIT89
contains a variety of input choices for the user pertaining to each evacuation simulation:
• Shortest travel route or user-defined route for occupants
• The use of CFAST smoke data, user-defined blockages, or none
• The choice of a body size for occupants which applies to entire population of the building - Large
(0.1458 m 2 ), Medium (0.1 13 m 2 ), or Small (0.0906 m 2 )
• The choice of speed for the entire population of the building - Emergency (horizontal unimpeded
speed = 1.36 m/s) or Normal (horizontal unimpeded speed = 0.91 m/s)
• A random delay time (uniform distribution)
• The modeling of disabled occupants, including the percentage of decrease in travel speed for
these occupants compared with the rest of the population
The three models described above were used in different capacities for this project. Simulex was used in
a limited capacity to provide analyses of occupant movement on individual floors and to simulate the
phased evacuation. EXIT89 and buildingEXODUS were used to model scenarios of an entire
WTC Tower with a large number of occupants within the building.
The three models were chosen for this project for different reasons. The Simulex model was chosen to
perform certain aspects of the project due to its heavy focus on understanding and accurately predicting
occupant movement, and was used for some of the smaller aspects of the project. The following reasons
justified buildingEXODUS and EXIT89 as reasonable choices to model WTC 1 and WTC 2:
• Both had the capacity to simulate evacuation from a 1 10-story building with 25,500 people.
• Both provided the output that was needed to answer evacuation questions, including visualization
capabilities from buildingEXODUS
• Both could handle the complexity of the building, such as the transfer floors and differences in
floor to ceiling heights in different part of the Tower
• Both provided explanation for the underlying data used for the movement of occupants
throughout the Tower
246 NIST NCSTAR 1 - 7, WTC Investigation
Egress Modeling
Both provided sufficient knowledge and direction to NIST, through personal contact and/or user
manuals, on how to use the model
•
• Both developers had published validation studies on the model
• Both models had been used by NIST or outside of NIST. This way NIST could identify their
inefficiencies ahead of time and compensate for these during this project.
D.4.1 Phased Evacuation (Scenario 1) Inputs and Results
All three models were used to perform the phased evacuation analysis, Scenario 1. Each model required
building geometry, population, and behavioral input conditions. The phased evacuation scenario involved
the simulation of floors 45-49, and the emergency floor was floor 48. All three stairs, containing 19 risers
per floor from floors 45-49, were added to the models with the appropriate diagonal length, stair width,
and door width into and out of the stairwell. The number of occupants used in the phased evacuation was
200 per floor on floors 47-49, which totaled to 600 occupants. The number of six hundred occupants was
used as a rounded estimate of the number of occupants seen on three floors within Zone 2 of the building.
The movement response delay was set to zero for one set of simulations, and ranged between zero and ten
minutes for the other set of simulations.
The difference between models for the phased evacuation scenario was the population characteristics of
the occupants. The population type used in the Simulex model was the "office staff type which
distributed 60 percent males and 40 percent females throughout the building. The population distribution
input into buildingEXODUS for the phased evacuation scenario was taken from WTC 1 telephone survey
demographic data: 5 percent males age 17-29, 38 percent Males age 30-50, 21 percent Males age 51-80,
3 percent Females age 17-29, 22 percent Females age 30-50, and 11 percent Females age 51-80. All 600
occupants simulated with the EXIT89 model were assigned the medium body size and emergency speed
parameters.
The total evacuation time for the phased evacuation, Scenario 1, was similar across all three models.
Without a time delay, Simulex predicted that occupants would travel to floor 45 in approximately 240
seconds, buildingEXODUS predicted 243 seconds, and EXIT89 (emergency conditions) predicted 210
seconds. With a to 10 minute time delay, Simulex predicted that occupants would travel to floor 45 in
approximately 690 seconds, buildingEXODUS predicted 660 seconds, and EXIT89 (emergency
conditions) predicted 690 seconds. These results are shown in Table D-4. From the three evacuation
models, the average minimum time to complete a phased evacuation in WTC 1 or WTC 2 building was
approximately 230 seconds, or around four minutes, assuming no evacuation delay and approximately
680 seconds, or around eleven minutes, assuming a to 10 minute time delay.
NIST NCSTAR 1-7, WTC Investigation 247
Appendix D
Table D-4. Phased evacuation model input conditions.
Evacuation Model
Occupant Type/ Characteristics
Evacuation
Time, No
Delay (s)
Evacuation
Time,
- 600 s
Delay (s)
Simulex
All office staff; 60 % men, 40% women
240
690
EXODUS
WTC 1 distribution: 5 % Males age 17-29, 38 % Males age
30-50, 21 % Males age 51-80, 3 % Females age 17-29, 22 %
Females age 30-50, and 1 1 % Females age 51-80
243
660
EXIT89
All Medium body size and emergency speed
210
690
D.4.2 Scenarios 2 and 3: Inputs and Results
Two models, buildingEXODUS and EXIT89, were used to perform the total building evacuation analysis
of a full-capacity tower with and without visitors, Scenarios 2 and 3. Each model required building
geometry, population, and behavioral input conditions.
buildingEXODUS
The buildingEXODUS model was used to simulate Scenarios 2 and 3, a total building evacuation of a
fully-occupied tower both with and without the inclusion of visitors. Without visitors, the population
totaled 19,800 people and with visitors, the population totaled 25,500 people. As discussed earlier, the
tower was input into the model using 1 1
representative floor plans from CAD, including
the four transfer floors. Mechanical floors were
input into the model; however, no occupants
were placed on these floors. All usable floor
space had to be filled with nodes measuring 0.5
m x 0.5 m on which the occupants would travel
throughout the building.
On each floor throughout the Tower, three
stairs were added in their appropriate position
on the floor and number of risers, depending
upon the floor. These stairs were added to the
model with the appropriate diagonal length,
stair width, door width into and out of the
stairwell, number of landings, and accurate stair
riser and tread distances.
Figure D-3 shows a representation of a stair
configuration of a 1.1 m (44 in.), 19 riser stair
(Stairs A or C) in the buildingEXODUS model.
The solid blue nodes represent the landing
space and the green patterned nodes represent
the stair steps. The figure shows a "spiral"
stair that is 2 nodes wide with the distances in
0.31 m on
diagonal of
stair nodes
0.6 m
between
centers of
landing nodes
Y
0.6 m between
centers of landing
nodes
Figure D-3. 44 in stair configuration in
buildingEXODUS.
248
NIST NCSTAR 1-7, WTC Investigation
Egress Modeling
between the centers of the nodes. The occupants can also move diagonally between the outer and inner
track of the stair, which is not shown in the figure to prevent confusion. Also, the dotted lines in between
the landing nodes are used to represent that one side of the landing will be located on one floor and the
other side of the landing will be located on the floor above. Due to the 7.5 in by 9.5 in riser and tread
configuration (not including the nose extending the tread space), the diagonal space between steps
measured approximately 12 in. Similar stair configuration was used for the 56 in. stair (Stair B). The
only difference was the distances from node to node; for instance, instead of a 0.6 m spacing between the
centers of landing nodes, there was a 0.7 m spacing input. The number of nodes specified along the width
of the stairwell is the buildingEXODUS's method of simulating stair width, and since not more than one
person is allowed to occupy the same node, specifying 2 nodes across each stairwell allows for 2-person
movement abreast in all three stairwells. NIST is aware of the narrowness of the 44 in. stair (Stair A and
C), however, did not find it appropriate to model only 1 -person abreast inside the stairwell throughout the
entire building. Therefore, two nodes were used across each stair as an assumption made when using the
buildingEXODUS model. Stair node connection lengths do not imply areas of the stairs, and therefore
density of the space, but rather the distance that occupants travel inside the stair.
For the population, either 19,800 occupants or 25,500 occupants (including visitors) were distributed
randomly throughout the building space, depending on the scenario. Similar to the phased evacuation
scenario, the following occupant types were used to describe the WTC tower population: 5 percent Males
age 17-29, 38 percent Males age 30-50, 21 percent Males age 51-80, 3 percent Females age 17-29,
22 percent Females age 30-50, and 1 1 percent Females age 51-80. All occupants were able-bodied
without mobility impairments.
The minimum evacuation time from the Tower was simulated by choosing an immediate response time
and the stair pack option, which predicted that occupants will "pack" inside the stairs during evacuation
(however in the simulation each occupant still tries to leave the step ahead of them empty). Also, it was
predicted that occupants traveled to the nearest stairwell on each floor. The simulation predicted that
occupants from a fully-occupied building without visitors, with a response delay of zero, evacuated the
building in approximately 110 minutes. Also, when visitors were added to the simulation,
buildingEXODUS predicted an evacuation time of 141 minutes.
Another group of simulations provided the evacuation time from the Tower if the occupants delayed their
evacuation by to 10 minutes. All other options chosen for the group of simulations were similar to
those described for the non delay simulations. buildingEXODUS predicted that occupants from a fully-
occupied building without visitors, with a response delay of to 10 minutes, evacuated the building in
approximately 112 minutes. Also, when visitors were added to the simulation, buildingEXODUS
predicted an evacuation time of 142 minutes.
EXIT89
The EXIT89 model was also used to simulate Scenarios 2 and 3, a total building evacuation of a fully-
occupied tower both with and without the inclusion of visitors. Without visitors, the population totaled
19,800 people and with visitors, the population totaled 25,500 people. The tower was input into the
model by describing each of the 1 1 representative floor plans with a series of nodes and arcs. On the
mechanical floors, only the stair nodes and arcs were associated with these floors, and no occupants were
NIST NCSTAR 1-7, WTC Investigation 249
Appendix D
placed on these floors. All usable floor space, including stairs and transfer corridors, was filled with
larger-sized nodes ranging from 2 m 2 to 104 m 2 and connected by arcs.
Similar to buildingEXODUS, three stairs were added in their appropriate position on each floor
throughout the tower. These stairs were added to the model with the appropriate diagonal length, stair
width, and door width into and out of the stairwell.
The stair input for the EXIT89 model was different than that used by the buildingEXODUS model.
Instead of the stair consisting of two lanes of nodes, a stairwell in EXIT89 was represented by a
horizontal area (m 2 ), stair width measured where stair nodes meet, and the stair travel distance measured
along the diagonal, which included the lengths traveled on the landings. Even though the stair input
method varied between buildingEXODUS and EXIT89, stair distances were verified as being similar in
both models for the Tower geometry.
Within EXIT89, the transfer corridors were input as horizontal nodes connecting vertical stair nodes at
each transfer floor, also part of the geometry input. At each transfer floor, the stairs leading to and
leading away from the floor were considered completely different stairwells and named accordingly.
However, since EXIT89 could account for only 10 different stairwells, the smallest transfer at floor 42
was considered part of a stairwell.
The number of occupants intentionally placed on each floor within the input file corresponded to the
occupant numbers calculated per zone from both the 14 m 2 /person and 1 1 m 2 /person densities, depending
upon the scenario. All occupants were assigned the medium body size and both emergency and normal
speed options (from multiple simulations as shown by the range of evacuation times in Table D-5).
The simulation predicted that occupants from a fully-occupied building without visitors, with a response
delay of zero, evacuated the building in approximately 97 to 117 minutes (emergency to normal speed
input conditions). Also, when visitors were added to the simulation, EXIT89 predicted an evacuation
time of 1 14 to 140 minutes.
When a response delay of 0-10 minutes was introduced to EXIT89, the simulation predicted that
occupants from a fully-occupied building without visitors evacuated the building in approximately 92 to
113 minutes. Also, when visitors were added to the simulation, EXIT89 predicted an evacuation time of
119 to 139 minutes.
250 NIST NCSTAR 1-7, WTC Investigation
Egress Modeling
Table D-5. Total Building Evacuation Time (Simulated) for Various Occupant Loads
Evacuation Model
Evacuation
Initiation Delay
Input
Evacuation Time
(min): 8,800
occupants
Evacuation time
(min): 19,800
occupants
Evacuation time
(min): 25,500
occupants
EXODUS
0- 10 Minute Delay
55
112
142
No Delay
52
110
141
EXIT89
0- 10 Minute Delay
71-74
92-113
119-139
No Delay
58-78
97 69 -117
114-140
D.4.3 Scenario Four (September 11th Capacity) Inputs and Results
The buildingEXODUS and EXIT89 models were used to perform the total building evacuation analysis
of a WTC tower, Scenario 4, with a population similar to that observed on the morning of
September 11, 2001. All inputs for the geometry and simulation options in buildingEXODUS and
EXIT89 were identical to those used in Scenarios 2 and 3. Also, the inputs for the population options
were identical, except for the number of occupants simulated. For scenario four, a total building
evacuation was simulated with 8,800 occupants (the capacity estimated for a tower on the morning of
September 11,2001).
buildingEXODUS predicted that occupants from a building holding a September 11, 2001, capacity
evacuated the building in approximately 52 minutes with no delay and 55 minutes with a to 10 minute
delay time. EXIT89 predicted that occupants from a building holding a September 11, 2001, capacity
evacuated the building in approximately 58 to 78 minutes (depending up the speed option chosen,
emergency to normal) with no delay and 71 to 74 minutes with a to 10 minute delay time.
D.4.4 Scenarios Five and Six (Full-Capacity with Damage) Inputs and Results
To model how the event would have changed if the buildings were fully occupied, models were run using
the calculated fully-occupied number of occupants in each tower, 19,800 occupants (without visitors).
Through running these scenarios, NIST hoped to answer the following questions:
• What would have been the consequences if WTC 1 had been fully-occupied on September 11,
2001, the building had been hit by aircraft, and then had collapsed 103 minutes later?
• What would have been the consequences if WTC 2 had been fully-occupied on September 11,
2001, the building had been hit by aircraft 16 minutes after WTC 1 was hit, and then had
collapsed 72 minutes after WTC 1 was hit (56 minutes from the time that WTC 2 was hit)?
To model these scenarios, NIST ran a series of calibration simulations for each tower in an attempt to
crudely simulate the occupant conditions inside the towers on September 11, 2001 with the capacity of
The underlying theory for people movement in EXIT89 is based upon the work of Predtechenskii and Milinskii. They
observed an inverse relationship between density and speed for three different types of movement: emergency, normal, and
comfortable. The effect of crowding (density) on overall evacuation speed was greater when no delay time was assumed. A
delay time when 19,800 occupants were present, according to EXIT89, spaces the occupants out and increases overall
evacuation efficiency. This was not found to be the case for all occupant loads (25,500).
NISTNCSTAR 1-7, WTC Investigation
251
Appendix D
8,800 occupants. Once the calibration simulations approximately achieved the goal of simulating the
September 11, 2001 scenario, the fully-occupied population was added to each tower. From the fully-
occupied simulations of each tower, possible consequences were predicted, such as the number of
occupants who would be trapped inside the tower at the time of collapse and the congestion points (the
Mezzanine escalator) that developed with a larger population.
For the calibration simulations, there were some difficult aspects of the evacuation that needed to be
accounted for in the evacuation, such as pre-evacuation activities, occupants changing stairs for various
reasons and the activities that they perform on the floor during evacuation, occupants resting in the stairs,
occupants being delayed by helping behaviors, occupants giving way to injured persons (superflow), and
firefighters counterflow. Although it was recognized that the models cannot accurately simulate all of
these activities as of yet, an attempt was made to account for time lost to perform such activities in the
calibration simulations. The buildingEXODUS model was the primary model used for this set of
scenarios. Due to the way that the stairs are created in the model (lanes of occupants instead of an
occupied space), it was difficult to model actual firefighters walking throughout the building and achieve
the same kind of occupant response and effect that were seen on September 11, 2001. It was also difficult
to simulate the effect of a larger bodied individual on the rest of the occupants in the stair, due to the
method of simulating stair movement used in the buildingEXODUS model (2-lane stairways).
The successful calibration simulation used for the WTC involved a method (to be referred to as) the "stop
and go" method. The stop and go method involved the creation of "doors" inside the stair that would
open and close throughout the simulation. This method attempted to account for evacuation interruptions
and the "stop and go" reports from survivors during the evacuation at different points throughout the stair,
such as leaving/changing stairs (sometimes occupants did this multiple times) for various reasons, resting
on the stairs, helping, waiting behind larger or disabled occupants, superflow, firefighter counterflow, etc.
Evacuation delays were distributed uniformly throughout the building, depending upon the Tower, to
account for pre-evacuation delays and any actions that occupants took when leaving the stairs, such as
short activities like moving directly to the next stairs, and longer activities such as seeking information,
making phone calls, looking for staff/coworkers, and looking out windows. In each building, a minimum
evacuation delay of 6 minutes was assigned to the population because even if the occupant left their floor
immediately, they were still faced with other challenges/actions taken while outside of the stairs.
Although the evacuation delay distribution on September 1 1 , 200 1 , was not uniform, this assumption was
made for the model.
One must remember that the "stop and go" method was only an estimate of the reported "stop and go"
activities observed in the WTC towers. This "stop and go" method alleviated a large crowd surrounding
the escalator, which was a result in the simulations run for Scenarios 2-4, but was not overwhelmingly
reported by survivors of September 11, 2001.
D.4.5 WTC 1 Calibration Simulation
The inputs for the calibration simulation of WTC 1 were the following:
• 7,200 occupants distributed from floors 90-ground floor
252 NIST NCSTAR 1-7, WTC Investigation
Egress Modeling
"Stop and go doors" located inside the stairwells at floor 76 (early in the evacuation) to simulate
the locked doors in the transfer, floor 5 (late in the evacuation), and floors 15 and 48 during the
evacuation
•
• 6-30 minute delay uniformly distributed over the entire population
• Distribution of gender and age used in Scenarios 2-4
• Stair packing and local potential
In the calibration simulation, 7,200 occupants were distributed from floor 90 to the ground floor. The
"stopping" points occurred (door closed) at every 5 minute interval for 2 minutes at floors 15 and 48
starting after 10 minutes. Then, at floor 76, the "door closed" only three times for 2 minutes each, to
simulate the locked doors on floor 76 in Stair A at the beginning of the simulation. Lastly, on floor 5, the
door closed 5 times for 2 minutes each at the ending time of the evacuation. Also, a 6 - 30 minute delay
was uniformly distributed to the entire population of the building.
The simulation was judged to be "successful" at predicting a September 11,2001, scenario if the
following criteria set for the WTC 1 calibration simulation were met:
• Evacuated more than 90 percent but less than 100 percent of the occupants from the tower before
WTC 2 collapse (72 min)
• Alleviated significant crowding at the entrance to the Mezzanine escalator
For the calibration simulation, when 7,200 occupants in WTC 1 were given a 6-30 minute time delay with
stoppage points, no significant queue developed at the escalator. Also, 7,200 occupants (under a uniform
delay) evacuated the building in approximately 71 minutes. This is a good estimate for WTC 1, because
90 percent of the occupants reported leaving the WTC complex before WTC 2 collapsed. Therefore, this
geometry and scenario set up was used to simulate a fully-occupied WTC 1, which contained
16,000 occupants below floor 92 (19,800 occupants minus those trapped above floor 91). Of the
remaining 1 percent who were trying to evacuate, however, some were located inside the towers at the
very base of the building (which the model falls a bit short of predicting), and some were located in the
Concourse area trying to evacuate the complex at this time. Further, this simulation did not account for
those who waited until WTC 2 collapsed to initiate their evacuation, (e.g., occupants on floor 64, and
some very high in the building trapped by significant floor damage).
D.4.6 WTC 1 Full-Capacity with Damage Simulation
When 16,000 occupants were simulated using the calibrated input assumptions, buildingEXODUS
calculated that there were still 2,400 occupants left inside WTC 1 as the building began to collapse.
There were larger queues simulated at the escalator in a full-capacity scenario when compared to the
September 11, 2001, calibration simulation. It is understood that with a larger population, occupants
would have encountered even more stoppages in the stairs than occupants did on September 11, 2001;
however, this is not quantifiable and only the delay times and stoppages used in the calibration simulation
were used in this scenario. Stoppage points in the full-capacity simulation continued until 10 minutes
after the collapse of WTC 2.
NIST NCSTAR 1-7, WTC Investigation 253
Appendix D
D.4.7 WTC 2 Calibration Simulation
The inputs for the calibration simulation of WTC 2 were the following:
7,400 occupants distributed from floors 107-ground floor (after subtracting elevator users).
"Stop and go doors" located inside the stairwells at floor 5 (late in the evacuation), and floor 15
and 48 during the evacuation, beginning after 20 minutes
"Door" closed inside stairs at floor 78 after 16 minutes
2-17 minute delay uniformly distributed over population from floors 107-78; 6-30 minute delay
uniformly distributed over the population from floors 77 - ground
Distribution of gender and age used in Scenarios 2-4
Stair packing and local potential
A total of 7,400 occupants (84 percent of 8,800) were specifically placed throughout each of the three
building sections, depending upon the percentage of occupants reporting elevator usage within that
section. Stoppage points were placed low in the building to simulate stopping for fire officials,
superflow, helping behaviors, etc. on floors 15 and 48 (throughout the evacuation beginning at
20 minutes) and floor 5 (toward the end of the simulation). The reason that the stoppages began after
20 minutes is because the fire department and others did not respond to WTC 2 until the building was hit.
This simulation modeled evacuation from the entire building until WTC 2 was attacked. After
16 minutes, it was assumed that everyone above the impact floor, 78, would not be able to evacuate past
the impact point, even though it is known that 1 8 occupants escaped from this area. The reason for
assigning a difference in delay times to occupants above and below the impact zone was because the time
delay was used to account for both pre-evacuation activities and activities performed during the
evacuation, such as waiting on a floor. Since occupants above the impact zone were not able to complete
their evacuation, a lower delay range was assigned to them.
The simulation was judged to be "successful" at predicting a September 11, 2001, scenario if the
following criteria set for the WTC 2 calibration simulation were met:
• All occupants below the 78th floor evacuated before T2 collapses (72 min)
• A loss of approximately 600-700 occupants was predicted above 78 when WTC 2 collapsed
• Alleviated significant crowding at the entrance to the Mezzanine escalator
The WTC 2 calibration simulation predicted that all occupants evacuate from below floor 78 and that
645 occupants remained above floor 78 when WTC 2 collapsed. Also, there was no significant queuing
predicted at the entrance to the Mezzanine escalator. These modeling assumptions were then used to
simulate the full-capacity tower containing 17,260 occupants, which accounted for occupants who used
elevators, as described below.
254 NIST NCSTAR 1-7, WTC Investigation
Egress Modeling
Elevators
For WTC 2, a fully occupied building on September 11, 2001 was assumed to contain 19,800 occupants.
It was necessary to estimate the number of occupants who would have been able to use the elevators
within the first 16 minutes on September 11, 2001, when WTC 2 was fully occupied. Calculations were
performed using the ELVAC model (Klote 1993). From the estimations made by the face-to-face
interview respondents using the shuttle elevators (NIST 2004), the peak speed of the shuttle elevators was
approximately 4.8 m/s (close to 78 floors in 60 seconds). The ELVAC model was used to calculate how
many occupants could have evacuated to the ground floor of WTC 2 in less than 1 6 minutes using
80 percent of the shuttle elevators available on floors 44 and 78. Only 80 percent of the elevators were
modeled, under the assumption that some elevators would be out of service. Under the assumptions that
the peak elevator speed was 4.8 m/s and had an acceleration of 1.5 m/s 2 , 20 percent of the occupants
(1,337) from the high section of the building (floors 78-110), 25 percent of the occupants (1,661) from the
middle section of the building (floors 44-77), and 6 percent of the occupants (388) from the lower section
of the building would have been able to evacuate via elevators before their tower was hit.
The maximum number of occupants who would have been able to use the shuttle elevators, which had a
capacity of 50 occupants each, to evacuate the building in under 14 minutes was estimated to be
3,386 occupants. However, since it was discovered from face-to-face interviews (NIST 2004) that some
elevators traveled to the ground floor without full capacity, it was assumed that only 75 percent of the
maximum number of occupants who could use elevators (0.75 x 3,386 = 2,540) evacuated using the
elevators. Therefore, approximately 17,260 occupants used the stairs in this simulation (19,800 - 2,540 =
17,260).
D.4.8 WTC 2 Full-Capacity with Damage Simulation
When 17,260 occupants were placed within the calibration simulation, buildingEXODUS predicted that
there were still over 8,000 occupants left inside WTC 2 after 72 minutes. Occupants were distributed
throughout the building, above and below the floors of impact, according to predicted elevator use, and
assigned an appropriate response delay. As was found with the WTC 1 full-capacity simulation, a larger
amount of queuing was predicted by the buildingEXODUS model, especially at the entrance to the
Mezzanine escalator.
Table D-6 shows the results for the calibration and fully-occupied scenarios with building damage for
both WTC 1 and WTC 2.
NISTNCSTAR 1-7, WTC Investigation 255
Appendix D
Table D-6. WTC 1 and WTC 2 full capacity evacuation simulation results.
Simulation
Title
Geometry;
Population
Simulation
# occupants evacuated
before tower collapses
WTC 1
Calibration 9/11
Floors 90-ground;
7,200 occupants
Response time = 6-30 min, stoppage
points throughout the stairs to simulate
stopping of occupants for approx. 2
minute intervals
7,200 evacuate from below
impact region within 71
minutes
WTC 1 fully-
occupied
building on 9/1 1
Floor 90-ground;
1 6,000 occupants
Response time = 6-30 min, stoppage
points throughout the stairs to simulate
stopping of occupants for approx. 2
minute intervals
13,600 occupants evacuated
before WTC 1 collapsed;
85 % evacuated from below
impact zone
WTC 2
Calibration 9/11
Floors 107-ground;
7,400 occupants
when removing
occupants using
elevators
Response time 2-17 minutes above 78;
6-30 minutes all other places; stoppage
points throughout the stairs that begin
after 20 minutes, "door" closes inside
stair at 78 after 16 minutes
6,755 occupants evacuated
before WTC 2 collapsed;
645 remained above floor
78
WTC 2 fully-
occupied
building on 9/1 1
Floors 107-ground;
1 7,260 occupants
when removing
occupants using
elevators
Response time = 6-30 min, stoppage
points throughout the stairs to simulate
stopping of occupants for approx. 2
minute intervals
8,883 occupants evacuate
before WTC 2 collapsed;
3,900 remained above floor
78; 58% evacuated the
WTC 2
Differences were found between the results of the full-capacity with damage simulations for WTC 1 and
WTC 2. The reasons for the differences in results are the following:
• In WTC 1, all "stop and go" points within each stairwell were ceased 10 minutes after the
collapse of WTC 2. This allowed "free flow" from 82 minutes to 103 minutes in WTC 1. In
WTC 2, the firefighters had no prior warning to stop counterflow measures and begin to evacuate.
Therefore, the "stop and go" points continued until the collapse of WTC 2.
• Occupants in WTC 1 had a total of 103 minutes before their building collapsed, whereas
occupants in WTC 2 had only 72 minutes.
• In WTC 2, with a much higher population, there were a greater number of congestion points
throughout the building, in addition to the "stop and go" points added to the building.
Additionally, some occupants higher in the building were given a longer evacuation time delay to
account for activities performed during evacuation, causing them to take longer than the allotted
72 minutes to reach the ground floor. These congestion points were located at the transfer floors.
• Only 58 percent of the occupants in the fully-occupied building evacuated WTC 2, via the use of
stairs or elevators. 2,540 were removed from the evacuation because they used elevators and
8,883 occupants evacuated via the stairs from WTC 2, for a total of 1 1,423 out of 19,800
occupants.
256
NIST NCSTAR 1-7, WTC Investigation
Egress Modeling
D.5
CONCLUSIONS AND FINDINGS
As shown in Table D-7, four different scenarios were run using three evacuation models, with and
without a 0-10 minute delay time randomly distributed among the occupants. The first column of results
shows the times for phased evacuation (with and without a delay time of 0-10 minutes) and the last three
columns show full evacuations of occupants with different building populations. On average, the phased
evacuation time (with initiation delays) for three floors to evacuate to three floors below the fire floor was
1 1 minutes. However, for a fully-occupied building (19,800 occupants), evacuation times were calculated
as 8 to 10 times larger for a total evacuation when compared to phased evacuation for this building.
These scenarios assume a to 10 minute delay time and do not account for other types of behavior that
may delay evacuation and increase total evacuation time.
Table D-7
. Summary ol
1 egress simulat
on results.
Evacuation
Model
Evacuation
Initiation Delay
Input
Phased
Evacuation:
600 occupants
Evacuation
Time (min):
8,800 occupants
Evacuation time
(min): 19,800
occupants
Evacuation time
(min): 25,500
occupants
EXODUS
0-10 Minute Delay
11
55
112
142
No Delay
4
52
110
141
EXIT89
0-10 Minute Delay
11.5
71-74
92-113
119-139
No Delay
3.5
58-78
97 70 -117
114-140
Simulex
0-10 Minute Delay
11.5
No Delay
4
Some mid-rise buildings that plan for phased evacuation can completely evacuate their building, at full
capacity, in less than 20 minutes. However, a building such as a World Trade Center tower produces full
evacuation times of over 100 minutes, which is a large difference in comparison to the phased evacuation
time of 1 1 minutes. With buildingEXODUS, adding a 0-10 minute delay did not significantly affect the
evacuation times from the building, adding only two additional minutes. The movement inside and from
the base of the stairwell was the limiting factor and controlled the evacuation time. When the starting
population roughly doubled, the evacuation time roughly doubled. As this simulation did not account for
behavioral aspects of the evacuation (such as leaving the stairs and resting), the full impact may be
underestimated somewhat by these evacuation model simulations.
Scenarios 2-4 estimated the total evacuation time for a fully-occupied building on September 11, 2001.
Figure D-4 shows that the evacuation with 25,585 occupants took approximately 2.5 times longer to
evacuate than the 8,800 occupant population. Therefore, to estimate the total evacuation time from a
fully-occupied building under the conditions observed on September 11, 2001, multiply by 2.5 the
observed time required to evacuate WTC 1 on September 1 1 (approximately 100 minutes). This estimate
is subject to the assumption that upon adding additional occupants, the emergency evacuation time scales
1 The underlying theory for people movement in EXIT89 is based upon the work of Predtechenskii and Milinskii. They
observed an inverse relationship between density and speed for three different types of movement: emergency, normal, and
comfortable. The effect of crowding (density) on overall evacuation speed was greater when no delay time was assumed. A
delay time when 19,800 occupants were present, according to EXIT89, spaces the occupants out and increases overall
evacuation efficiency. This was not found to be the case for all occupant loads (25,500).
NISTNCSTAR 1-7, WTC Investigation
257
Appendix D
linearly, as was observed with the modeled evacuation time. A fully occupied building with visitors, but
without elevator usage, would have required approximately four hours (2.5 x 100 minutes « 4 hours ) to
evacuate the building. Using the same method, if WTC 1 was occupied with approximately 20,000
occupants (without visitors), the evacuation would have required roughly twice as long to evacuate than
the 8,800 occupant simulation. Therefore, a fully occupied building without visitors and without the use
of elevators would have required over three hours (2 x 1 00 minutes « 3 hours ) to evacuate the building.
Note that in 2005, PANYNJ estimated that the maximum population of WTC 1 or WTC 2 would not
likely have exceeded 20,000. 71
Full Building Evacuation of a WTC Tower
160
140
c
120
E
a>
100
b
H
c
80
o
(8
3
60
o
cB
>
HI
40
20
25,585 - full capacity
with visitors
19,800 - full capacity
8,800 - 9/11 capacity
5000 10000 15000 20000
Number of occupants in Tower
25000
30000
Figure D-4. Full building evacuation simulation results with differing building
populations.
Scenarios 5-6 were modeled to estimate the consequences of a fully-occupied building on
September 11, 2001 by estimating the number of lives lost at each building collapse. As the results show,
more people would have lost their lives, particularly in WTC 2 due to its earlier collapse. Table D-8
shows the simulated number of occupants (by starting location) who successfully evacuated WTC 1 and
WTC 2 for a fully-occupied building subject to the conditions observed on September 11, 2001.
• WTC 1-69 percent of all occupants (13,600 occupants out of 19,800) would have successfully
evacuated the Tower, which breaks down to percent (0 out of 3,800) from above the impact
zone and 85 percent (13,600 out of 16,000) from below the impact zone in 103 minutes.
Bhol, Saroj. PANYNJ (September 21, 2005). Email from S. Bhol to S. Sunder in response to NIST question.
258
NIST NCSTAR 1-7, WTC Investigation
Egress Modeling
• WTC 2-58 percent of all occupants (8,883 by stairs and 2,540 by elevators) would have
successfully evacuated the tower with the use of stairs and elevators in 72 minutes, which breaks
down to 44 percent from above the impact zone and 75 percent from below the impact zone in 72
minutes.
Table D-8. Simulated evacuation results for fully-occupied WTC 1 and WTC 2 on
September 11, 2001.
Building
Total
Number of
Occupants
at t = 0.0
Potential
Number
of
Evacuees
Occupants
Remaining
in Building
at Collapse
Occupants
Trapped
Above
Floors of
Impact
Percentage of Occupants Who
Successfully Evacuated Relative to
Where They Started
Total
Below
Impact
Above
Impact
WTC 1
19,800
16,000
6,200
3,800
69%
85%
0%
WTC 2
19,800
17,260
8,377
3,900
58%
75%
44%
Total
39,600
33,260
14,577
7,700
63%
NISTNCSTAR 1-7, WTC Investigation
259
Appendix D
This page intentionally left blank.
260 NIST NCSTAR 1-7, WTC Investigation
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Open Library
American Libraries
TV News
Understanding 9/11